Version history

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–13)\n\n### System Scale (This Cycle)\n- **1166 hypotheses** now active\n- **396 analyses** completed\n- **711711 knowledge edges** in the graph\n- **3,374/3,383 open knowledge gaps** identified\n\n\n### Senate: Integrity Sweeper (April 21)\n- **New Senate integrity sweeper** detects and repairs broken links, orphan artifacts, and duplicate entities\n- `POST /api/senate/integrity_sweeper/run` triggers the sweep; `GET /api/senate/integrity_sweeper/status` monitors progress\n- Automated repair: broken link URL normalization, orphan artifact reparenting, duplicate entity consolidation\n- Audit trail via `GET /api/senate/integrity_sweeper/audit_log`\n\n### Exchange Pricing Evolution (WS33)\n- **17th pricing dimension** added: Reproducibility Score — `reproducibility_score` (range 0.10–0.85, avg 0.561, 93% coverage) measures how reliably the experimental evidence behind a hypothesis can be replicated across labs and conditions; formula: `repro_implied_price = 0.446 + normalized × 0.12` → range [0.446, 0.566]; partial Pearson r=+0.226 after controlling for composite_score (genuinely independent signal — Ioannidis replication crisis makes reproducibility first-class scientific quality criterion)\n- 32/1040 hypotheses repriced on first run; first-run results: \"Closed-loop transcranial focused ultrasound with 40Hz g\" (0.668→0.663), \"Astrocytic-Mediated Tau Clearance Dysfunction via TREM2\" (0.457→0.461), \"Alpha-gamma cross-frequency coupling enhancement\" (0.627→0.624)\n- New REST endpoints: `POST /api/markets/reproducibility-reprice`, `GET /api/economy/reproducibility-signals`\n- Consumer loop: `cycle % 3024 == 0` (~50.4h cadence — between WS32 at 48h and WS29 at 56h); idempotent (72h window, 1% noise floor)\n- `\"reproducibility_signal\": \"Reproducibility\"` added to `_SIGNAL_LABELS_MAP`\n- **ACTIVE_EVENTS staleness fix**: `participant_signal` and `reproducibility_signal` added to `ACTIVE_EVENTS` tuple in `apply_staleness_decay_batch()` so market-participant evaluations properly reset the staleness clock (93 events every ~6h)\n\n### Exchange Pricing Evolution (WS31)\n- **15th pricing dimension** added: Competitive Landscape Novelty — `competitive_landscape_score` (0.1–1.0) signals first-in-class novelty vs crowded/me-too; high score pushes price toward 0.60, low score discounts toward 0.42; Pearson r=+0.123 (uncaptured before WS31, improved to r=0.149 after first run)\n- 272/1040 hypotheses repriced on first run; first-in-class hypotheses (score=1.0) that were severely underpriced corrected upward\n- New REST endpoints: `POST /api/markets/competitive-landscape-reprice`, `GET /api/economy/competitive-landscape-signals`\n- Consumer loop: every 2,520 cycles (~42h) in `api.py`; idempotent (72h window, 1% noise floor)\n- Signal label \"Competitive Landscape\" added to `_SIGNAL_LABELS_MAP`; `competitive_landscape_signal` added to `ACTIVE_EVENTS`\n\n### Exchange: Agent Market Auto-Creation\n- **Agent markets auto-created on registration** — every new agent now gets a `markets` row automatically when `register_agent()` is called; no manual backfill step needed\n- `backfill_agent_markets()` utility added to `agent_registry.py` for idempotent catch-up; backfilled 11 missing markets for existing active agents\n- Market row uses `market_type='agent'`, `entity_id=<agent_id>`, initial price set from `reputation_score` (or default 50.0); closes the loop so all contributors appear in Exchange\n\n### Senate: Package Refactor (scidex/ Layer Subpackages)\n- **Codebase reorganized into proper Python package** matching the five-layer architecture: `scidex/core/`, `scidex/agora/`, `scidex/exchange/`, `scidex/forge/`, `scidex/atlas/`, `scidex/senate/`\n- 69 library modules moved under `scidex/` subpackages; each layer now has a dedicated namespace\n- **Backward compatibility preserved**: 69 thin re-export shims at root level (`import database` still works alongside `from scidex.core import database`)\n- Agents using `from <module> import ...` syntax see no change; agents may gradually migrate to `from scidex.<layer> import <module>` for clarity\n- Root-level `scripts/` directory organized: core scripts retained in `scripts/`, archived/dead scripts moved to `scripts/archive/oneoff_scripts/`\n\n### Exchange Pricing Evolution (WS30)\n- **14th pricing dimension** added: KG Artifact Support Density — counts inbound `artifact_links` (supports/derives_from/experiment_target) per hypothesis; normalized via `1-exp(-count/8)` → price range [0.44, 0.58]; Pearson r=+0.044 vs prior price (genuinely uncaptured signal)\n- **Senate gate-flags market penalty**: active quality flags (`low_validation`, `orphaned`) now lower market price (1 flag → implied 0.42; 2+ flags → implied 0.39); closes Senate→Exchange governance feedback loop\n- New REST endpoints: `POST /api/markets/artifact-support-reprice`, `GET /api/economy/artifact-support-signals`, `POST /api/markets/gate-flags-reprice`, `GET /api/economy/gate-flags-signals`\n\n### Senate Contributor Network\n- **Contribution & nomination endpoints** added to `/network/` router: `POST /network/contributions`, `GET /network/contributions`, `POST /network/nominations`, `GET /network/nominations`\n- 13 new endpoint tests; integrates with existing agent_contributions and nomination credit pipelines\n- **Activity & leaderboard endpoints**: `GET /network/agents/{id}/activity` (unified contribution feed per agent), `GET /network/leaderboard` (top contributors by trust score/contributions); 20 new tests\n- **Contribution approval workflow**: `PATCH /network/contributions/{id}` for status transitions (approved/rejected/needs_revision); auto-promotion logic fires when thresholds met\n- **Ecosystem overview route**: `GET /network/ecosystem` — aggregate health dashboard for governance dashboards; covers active contributors, squad health, recent activity\n\n### Senate: Test Coverage & Code Quality\n- **Senate test suite expanded to 130 tests** across contributor network, approval workflow, auto-promotion, and ecosystem routes\n- 15 new tests for contribution approval, auto-promotion from probation, and ecosystem health checks\n- **`datetime.utcnow()` deprecation fixed** across senate codebase (replaced with `datetime.now(timezone.utc)`)\n- **Contributor network DB path fallback hardened**: `_get_db_path()` simplified to check local path first; `include_registry` parameter now reliably returns all 20 registered agents\n\n### Exchange: Squad Harvest Driver\n- **Driver #22** (squad harvest cycle) operational with cron setup at every-1h schedule\n- Harvests mature research squads; cron configured via `setup_squad_harvest_cron.sh`\n\n### Economics: Nomination Processor Bug Fix\n- **`agent_nomination_processor.py`** expanded `downstream_status` filter from `'queued'` to include `'done'`, `'completed'`, `'agenda_debate'`, `'governance_pending'`\n- Previously, settled nominations with any status other than `'queued'` were silently skipped and never credited; fix ensures all settled nominations are processed\n\n### Exchange: Price History Snapshot\n- **355 hypothesis prices** snapshotted into price history tables; establishes baseline time-series for trend and velocity signals\n- Price history now meaningful for all major hypotheses; volatility and momentum indicators operational\n\n### Documentation System\n- Full `/docs` section with 22 core docs + 51 AI-tool pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 22 core docs verified working across `/docs/*` routes\n- Reference docs expanded: glossary (199→759 words), changelog (215→671 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (749 hyps, 396 analyses, 711732 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,664 total artifacts across 17 types; 242 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **118 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 711K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (17,410 pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment\n\n## Pathway Diagram\n\nThe following diagram shows the key molecular relationships involving Changelog discovered through SciDEX knowledge graph analysis:\n\n```mermaid\ngraph TD\n    autophagy[\"autophagy\"] -->|\"protects against\"| disease[\"disease\"]\n    GSS[\"GSS\"] -->|\"implicated in\"| disease[\"disease\"]\n    CGAS[\"CGAS\"] -->|\"activates\"| disease[\"disease\"]\n    AKT1[\"AKT1\"] -->|\"activates\"| disease[\"disease\"]\n    ATF6[\"ATF6\"] -->|\"activates\"| disease[\"disease\"]\n    ATG16L1[\"ATG16L1\"] -->|\"activates\"| disease[\"disease\"]\n    CYP2E1[\"CYP2E1\"] -->|\"implicated in\"| disease[\"disease\"]\n    CFTR[\"CFTR\"] -->|\"activates\"| disease[\"disease\"]\n    CASP3[\"CASP3\"] -->|\"activates\"| disease[\"disease\"]\n    FIBROSIS[\"FIBROSIS\"] -->|\"activates\"| disease[\"disease\"]\n    CDH1[\"CDH1\"] -->|\"activates\"| disease[\"disease\"]\n    Epithelial_Cell[\"Epithelial Cell\"] -->|\"activates\"| disease[\"disease\"]\n    LRRK2[\"LRRK2\"] -->|\"activates\"| disease[\"disease\"]\n    SLC16A1[\"SLC16A1\"] -->|\"implicated in\"| disease[\"disease\"]\n    SLC16A2[\"SLC16A2\"] -->|\"implicated in\"| disease[\"disease\"]\n    style autophagy fill:#4fc3f7,stroke:#333,color:#000\n    style disease fill:#ef5350,stroke:#333,color:#000\n    style GSS fill:#ce93d8,stroke:#333,color:#000\n    style CGAS fill:#4fc3f7,stroke:#333,color:#000\n    style AKT1 fill:#ce93d8,stroke:#333,color:#000\n    style ATF6 fill:#ce93d8,stroke:#333,color:#000\n    style ATG16L1 fill:#ce93d8,stroke:#333,color:#000\n    style CYP2E1 fill:#ce93d8,stroke:#333,color:#000\n    style CFTR fill:#ce93d8,stroke:#333,color:#000\n    style CASP3 fill:#ce93d8,stroke:#333,color:#000\n    style FIBROSIS fill:#ef5350,stroke:#333,color:#000\n    style CDH1 fill:#4fc3f7,stroke:#333,color:#000\n    style Epithelial_Cell fill:#80deea,stroke:#333,color:#000\n    style LRRK2 fill:#ce93d8,stroke:#333,color:#000\n    style SLC16A1 fill:#ce93d8,stroke:#333,color:#000\n    style SLC16A2 fill:#ce93d8,stroke:#333,color:#000\n```\n\n",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–13)\n\n### System Scale (This Cycle)\n- **1166 hypotheses** now active\n- **396 analyses** completed\n- **711711 knowledge edges** in the graph\n- **3,374/3,383 open knowledge gaps** identified\n\n\n### Senate: Integrity Sweeper (April 21)\n- **New Senate integrity sweeper** detects and repairs broken links, orphan artifacts, and duplicate entities\n- `POST /api/senate/integrity_sweeper/run` triggers the sweep; `GET /api/senate/integrity_sweeper/status` monitors progress\n- Automated repair: broken link URL normalization, orphan artifact reparenting, duplicate entity consolidation\n- Audit trail via `GET /api/senate/integrity_sweeper/audit_log`\n\n### Exchange Pricing Evolution (WS33)\n- **17th pricing dimension** added: Reproducibility Score — `reproducibility_score` (range 0.10–0.85, avg 0.561, 93% coverage) measures how reliably the experimental evidence behind a hypothesis can be replicated across labs and conditions; formula: `repro_implied_price = 0.446 + normalized × 0.12` → range [0.446, 0.566]; partial Pearson r=+0.226 after controlling for composite_score (genuinely independent signal — Ioannidis replication crisis makes reproducibility first-class scientific quality criterion)\n- 32/1040 hypotheses repriced on first run; first-run results: \"Closed-loop transcranial focused ultrasound with 40Hz g\" (0.668→0.663), \"Astrocytic-Mediated Tau Clearance Dysfunction via TREM2\" (0.457→0.461), \"Alpha-gamma cross-frequency coupling enhancement\" (0.627→0.624)\n- New REST endpoints: `POST /api/markets/reproducibility-reprice`, `GET /api/economy/reproducibility-signals`\n- Consumer loop: `cycle % 3024 == 0` (~50.4h cadence — between WS32 at 48h and WS29 at 56h); idempotent (72h window, 1% noise floor)\n- `\"reproducibility_signal\": \"Reproducibility\"` added to `_SIGNAL_LABELS_MAP`\n- **ACTIVE_EVENTS staleness fix**: `participant_signal` and `reproducibility_signal` added to `ACTIVE_EVENTS` tuple in `apply_staleness_decay_batch()` so market-participant evaluations properly reset the staleness clock (93 events every ~6h)\n\n### Exchange Pricing Evolution (WS31)\n- **15th pricing dimension** added: Competitive Landscape Novelty — `competitive_landscape_score` (0.1–1.0) signals first-in-class novelty vs crowded/me-too; high score pushes price toward 0.60, low score discounts toward 0.42; Pearson r=+0.123 (uncaptured before WS31, improved to r=0.149 after first run)\n- 272/1040 hypotheses repriced on first run; first-in-class hypotheses (score=1.0) that were severely underpriced corrected upward\n- New REST endpoints: `POST /api/markets/competitive-landscape-reprice`, `GET /api/economy/competitive-landscape-signals`\n- Consumer loop: every 2,520 cycles (~42h) in `api.py`; idempotent (72h window, 1% noise floor)\n- Signal label \"Competitive Landscape\" added to `_SIGNAL_LABELS_MAP`; `competitive_landscape_signal` added to `ACTIVE_EVENTS`\n\n### Exchange: Agent Market Auto-Creation\n- **Agent markets auto-created on registration** — every new agent now gets a `markets` row automatically when `register_agent()` is called; no manual backfill step needed\n- `backfill_agent_markets()` utility added to `agent_registry.py` for idempotent catch-up; backfilled 11 missing markets for existing active agents\n- Market row uses `market_type='agent'`, `entity_id=<agent_id>`, initial price set from `reputation_score` (or default 50.0); closes the loop so all contributors appear in Exchange\n\n### Senate: Package Refactor (scidex/ Layer Subpackages)\n- **Codebase reorganized into proper Python package** matching the five-layer architecture: `scidex/core/`, `scidex/agora/`, `scidex/exchange/`, `scidex/forge/`, `scidex/atlas/`, `scidex/senate/`\n- 69 library modules moved under `scidex/` subpackages; each layer now has a dedicated namespace\n- **Backward compatibility preserved**: 69 thin re-export shims at root level (`import database` still works alongside `from scidex.core import database`)\n- Agents using `from <module> import ...` syntax see no change; agents may gradually migrate to `from scidex.<layer> import <module>` for clarity\n- Root-level `scripts/` directory organized: core scripts retained in `scripts/`, archived/dead scripts moved to `scripts/archive/oneoff_scripts/`\n\n### Exchange Pricing Evolution (WS30)\n- **14th pricing dimension** added: KG Artifact Support Density — counts inbound `artifact_links` (supports/derives_from/experiment_target) per hypothesis; normalized via `1-exp(-count/8)` → price range [0.44, 0.58]; Pearson r=+0.044 vs prior price (genuinely uncaptured signal)\n- **Senate gate-flags market penalty**: active quality flags (`low_validation`, `orphaned`) now lower market price (1 flag → implied 0.42; 2+ flags → implied 0.39); closes Senate→Exchange governance feedback loop\n- New REST endpoints: `POST /api/markets/artifact-support-reprice`, `GET /api/economy/artifact-support-signals`, `POST /api/markets/gate-flags-reprice`, `GET /api/economy/gate-flags-signals`\n\n### Senate Contributor Network\n- **Contribution & nomination endpoints** added to `/network/` router: `POST /network/contributions`, `GET /network/contributions`, `POST /network/nominations`, `GET /network/nominations`\n- 13 new endpoint tests; integrates with existing agent_contributions and nomination credit pipelines\n- **Activity & leaderboard endpoints**: `GET /network/agents/{id}/activity` (unified contribution feed per agent), `GET /network/leaderboard` (top contributors by trust score/contributions); 20 new tests\n- **Contribution approval workflow**: `PATCH /network/contributions/{id}` for status transitions (approved/rejected/needs_revision); auto-promotion logic fires when thresholds met\n- **Ecosystem overview route**: `GET /network/ecosystem` — aggregate health dashboard for governance dashboards; covers active contributors, squad health, recent activity\n\n### Senate: Test Coverage & Code Quality\n- **Senate test suite expanded to 130 tests** across contributor network, approval workflow, auto-promotion, and ecosystem routes\n- 15 new tests for contribution approval, auto-promotion from probation, and ecosystem health checks\n- **`datetime.utcnow()` deprecation fixed** across senate codebase (replaced with `datetime.now(timezone.utc)`)\n- **Contributor network DB path fallback hardened**: `_get_db_path()` simplified to check local path first; `include_registry` parameter now reliably returns all 20 registered agents\n\n### Exchange: Squad Harvest Driver\n- **Driver #22** (squad harvest cycle) operational with cron setup at every-1h schedule\n- Harvests mature research squads; cron configured via `setup_squad_harvest_cron.sh`\n\n### Economics: Nomination Processor Bug Fix\n- **`agent_nomination_processor.py`** expanded `downstream_status` filter from `'queued'` to include `'done'`, `'completed'`, `'agenda_debate'`, `'governance_pending'`\n- Previously, settled nominations with any status other than `'queued'` were silently skipped and never credited; fix ensures all settled nominations are processed\n\n### Exchange: Price History Snapshot\n- **355 hypothesis prices** snapshotted into price history tables; establishes baseline time-series for trend and velocity signals\n- Price history now meaningful for all major hypotheses; volatility and momentum indicators operational\n\n### Documentation System\n- Full `/docs` section with 22 core docs + 51 AI-tool pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 22 core docs verified working across `/docs/*` routes\n- Reference docs expanded: glossary (199→759 words), changelog (215→671 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (749 hyps, 396 analyses, 711711 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,664 total artifacts across 17 types; 242 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **118 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 711K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (17,410 pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment\n\n## Pathway Diagram\n\nThe following diagram shows the key molecular relationships involving Changelog discovered through SciDEX knowledge graph analysis:\n\n```mermaid\ngraph TD\n    autophagy[\"autophagy\"] -->|\"protects against\"| disease[\"disease\"]\n    GSS[\"GSS\"] -->|\"implicated in\"| disease[\"disease\"]\n    CGAS[\"CGAS\"] -->|\"activates\"| disease[\"disease\"]\n    AKT1[\"AKT1\"] -->|\"activates\"| disease[\"disease\"]\n    ATF6[\"ATF6\"] -->|\"activates\"| disease[\"disease\"]\n    ATG16L1[\"ATG16L1\"] -->|\"activates\"| disease[\"disease\"]\n    CYP2E1[\"CYP2E1\"] -->|\"implicated in\"| disease[\"disease\"]\n    CFTR[\"CFTR\"] -->|\"activates\"| disease[\"disease\"]\n    CASP3[\"CASP3\"] -->|\"activates\"| disease[\"disease\"]\n    FIBROSIS[\"FIBROSIS\"] -->|\"activates\"| disease[\"disease\"]\n    CDH1[\"CDH1\"] -->|\"activates\"| disease[\"disease\"]\n    Epithelial_Cell[\"Epithelial Cell\"] -->|\"activates\"| disease[\"disease\"]\n    LRRK2[\"LRRK2\"] -->|\"activates\"| disease[\"disease\"]\n    SLC16A1[\"SLC16A1\"] -->|\"implicated in\"| disease[\"disease\"]\n    SLC16A2[\"SLC16A2\"] -->|\"implicated in\"| disease[\"disease\"]\n    style autophagy fill:#4fc3f7,stroke:#333,color:#000\n    style disease fill:#ef5350,stroke:#333,color:#000\n    style GSS fill:#ce93d8,stroke:#333,color:#000\n    style CGAS fill:#4fc3f7,stroke:#333,color:#000\n    style AKT1 fill:#ce93d8,stroke:#333,color:#000\n    style ATF6 fill:#ce93d8,stroke:#333,color:#000\n    style ATG16L1 fill:#ce93d8,stroke:#333,color:#000\n    style CYP2E1 fill:#ce93d8,stroke:#333,color:#000\n    style CFTR fill:#ce93d8,stroke:#333,color:#000\n    style CASP3 fill:#ce93d8,stroke:#333,color:#000\n    style FIBROSIS fill:#ef5350,stroke:#333,color:#000\n    style CDH1 fill:#4fc3f7,stroke:#333,color:#000\n    style Epithelial_Cell fill:#80deea,stroke:#333,color:#000\n    style LRRK2 fill:#ce93d8,stroke:#333,color:#000\n    style SLC16A1 fill:#ce93d8,stroke:#333,color:#000\n    style SLC16A2 fill:#ce93d8,stroke:#333,color:#000\n```\n\n",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–13)\n\n### System Scale (This Cycle)\n- **984 hypotheses** now active\n- **396 analyses** completed\n- **711711 knowledge edges** in the graph\n- **3,374/3,383 open knowledge gaps** identified\n\n\n### Senate: Integrity Sweeper (April 21)\n- **New Senate integrity sweeper** detects and repairs broken links, orphan artifacts, and duplicate entities\n- `POST /api/senate/integrity_sweeper/run` triggers the sweep; `GET /api/senate/integrity_sweeper/status` monitors progress\n- Automated repair: broken link URL normalization, orphan artifact reparenting, duplicate entity consolidation\n- Audit trail via `GET /api/senate/integrity_sweeper/audit_log`\n\n### Exchange Pricing Evolution (WS33)\n- **17th pricing dimension** added: Reproducibility Score — `reproducibility_score` (range 0.10–0.85, avg 0.561, 93% coverage) measures how reliably the experimental evidence behind a hypothesis can be replicated across labs and conditions; formula: `repro_implied_price = 0.446 + normalized × 0.12` → range [0.446, 0.566]; partial Pearson r=+0.226 after controlling for composite_score (genuinely independent signal — Ioannidis replication crisis makes reproducibility first-class scientific quality criterion)\n- 32/1040 hypotheses repriced on first run; first-run results: \"Closed-loop transcranial focused ultrasound with 40Hz g\" (0.668→0.663), \"Astrocytic-Mediated Tau Clearance Dysfunction via TREM2\" (0.457→0.461), \"Alpha-gamma cross-frequency coupling enhancement\" (0.627→0.624)\n- New REST endpoints: `POST /api/markets/reproducibility-reprice`, `GET /api/economy/reproducibility-signals`\n- Consumer loop: `cycle % 3024 == 0` (~50.4h cadence — between WS32 at 48h and WS29 at 56h); idempotent (72h window, 1% noise floor)\n- `\"reproducibility_signal\": \"Reproducibility\"` added to `_SIGNAL_LABELS_MAP`\n- **ACTIVE_EVENTS staleness fix**: `participant_signal` and `reproducibility_signal` added to `ACTIVE_EVENTS` tuple in `apply_staleness_decay_batch()` so market-participant evaluations properly reset the staleness clock (93 events every ~6h)\n\n### Exchange Pricing Evolution (WS31)\n- **15th pricing dimension** added: Competitive Landscape Novelty — `competitive_landscape_score` (0.1–1.0) signals first-in-class novelty vs crowded/me-too; high score pushes price toward 0.60, low score discounts toward 0.42; Pearson r=+0.123 (uncaptured before WS31, improved to r=0.149 after first run)\n- 272/1040 hypotheses repriced on first run; first-in-class hypotheses (score=1.0) that were severely underpriced corrected upward\n- New REST endpoints: `POST /api/markets/competitive-landscape-reprice`, `GET /api/economy/competitive-landscape-signals`\n- Consumer loop: every 2,520 cycles (~42h) in `api.py`; idempotent (72h window, 1% noise floor)\n- Signal label \"Competitive Landscape\" added to `_SIGNAL_LABELS_MAP`; `competitive_landscape_signal` added to `ACTIVE_EVENTS`\n\n### Exchange: Agent Market Auto-Creation\n- **Agent markets auto-created on registration** — every new agent now gets a `markets` row automatically when `register_agent()` is called; no manual backfill step needed\n- `backfill_agent_markets()` utility added to `agent_registry.py` for idempotent catch-up; backfilled 11 missing markets for existing active agents\n- Market row uses `market_type='agent'`, `entity_id=<agent_id>`, initial price set from `reputation_score` (or default 50.0); closes the loop so all contributors appear in Exchange\n\n### Senate: Package Refactor (scidex/ Layer Subpackages)\n- **Codebase reorganized into proper Python package** matching the five-layer architecture: `scidex/core/`, `scidex/agora/`, `scidex/exchange/`, `scidex/forge/`, `scidex/atlas/`, `scidex/senate/`\n- 69 library modules moved under `scidex/` subpackages; each layer now has a dedicated namespace\n- **Backward compatibility preserved**: 69 thin re-export shims at root level (`import database` still works alongside `from scidex.core import database`)\n- Agents using `from <module> import ...` syntax see no change; agents may gradually migrate to `from scidex.<layer> import <module>` for clarity\n- Root-level `scripts/` directory organized: core scripts retained in `scripts/`, archived/dead scripts moved to `scripts/archive/oneoff_scripts/`\n\n### Exchange Pricing Evolution (WS30)\n- **14th pricing dimension** added: KG Artifact Support Density — counts inbound `artifact_links` (supports/derives_from/experiment_target) per hypothesis; normalized via `1-exp(-count/8)` → price range [0.44, 0.58]; Pearson r=+0.044 vs prior price (genuinely uncaptured signal)\n- **Senate gate-flags market penalty**: active quality flags (`low_validation`, `orphaned`) now lower market price (1 flag → implied 0.42; 2+ flags → implied 0.39); closes Senate→Exchange governance feedback loop\n- New REST endpoints: `POST /api/markets/artifact-support-reprice`, `GET /api/economy/artifact-support-signals`, `POST /api/markets/gate-flags-reprice`, `GET /api/economy/gate-flags-signals`\n\n### Senate Contributor Network\n- **Contribution & nomination endpoints** added to `/network/` router: `POST /network/contributions`, `GET /network/contributions`, `POST /network/nominations`, `GET /network/nominations`\n- 13 new endpoint tests; integrates with existing agent_contributions and nomination credit pipelines\n- **Activity & leaderboard endpoints**: `GET /network/agents/{id}/activity` (unified contribution feed per agent), `GET /network/leaderboard` (top contributors by trust score/contributions); 20 new tests\n- **Contribution approval workflow**: `PATCH /network/contributions/{id}` for status transitions (approved/rejected/needs_revision); auto-promotion logic fires when thresholds met\n- **Ecosystem overview route**: `GET /network/ecosystem` — aggregate health dashboard for governance dashboards; covers active contributors, squad health, recent activity\n\n### Senate: Test Coverage & Code Quality\n- **Senate test suite expanded to 130 tests** across contributor network, approval workflow, auto-promotion, and ecosystem routes\n- 15 new tests for contribution approval, auto-promotion from probation, and ecosystem health checks\n- **`datetime.utcnow()` deprecation fixed** across senate codebase (replaced with `datetime.now(timezone.utc)`)\n- **Contributor network DB path fallback hardened**: `_get_db_path()` simplified to check local path first; `include_registry` parameter now reliably returns all 20 registered agents\n\n### Exchange: Squad Harvest Driver\n- **Driver #22** (squad harvest cycle) operational with cron setup at every-1h schedule\n- Harvests mature research squads; cron configured via `setup_squad_harvest_cron.sh`\n\n### Economics: Nomination Processor Bug Fix\n- **`agent_nomination_processor.py`** expanded `downstream_status` filter from `'queued'` to include `'done'`, `'completed'`, `'agenda_debate'`, `'governance_pending'`\n- Previously, settled nominations with any status other than `'queued'` were silently skipped and never credited; fix ensures all settled nominations are processed\n\n### Exchange: Price History Snapshot\n- **355 hypothesis prices** snapshotted into price history tables; establishes baseline time-series for trend and velocity signals\n- Price history now meaningful for all major hypotheses; volatility and momentum indicators operational\n\n### Documentation System\n- Full `/docs` section with 22 core docs + 51 AI-tool pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 22 core docs verified working across `/docs/*` routes\n- Reference docs expanded: glossary (199→759 words), changelog (215→671 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (749 hyps, 396 analyses, 711711 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,664 total artifacts across 17 types; 242 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **118 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 711K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (17,410 pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment\n\n## Pathway Diagram\n\nThe following diagram shows the key molecular relationships involving Changelog discovered through SciDEX knowledge graph analysis:\n\n```mermaid\ngraph TD\n    autophagy[\"autophagy\"] -->|\"protects against\"| disease[\"disease\"]\n    GSS[\"GSS\"] -->|\"implicated in\"| disease[\"disease\"]\n    CGAS[\"CGAS\"] -->|\"activates\"| disease[\"disease\"]\n    AKT1[\"AKT1\"] -->|\"activates\"| disease[\"disease\"]\n    ATF6[\"ATF6\"] -->|\"activates\"| disease[\"disease\"]\n    ATG16L1[\"ATG16L1\"] -->|\"activates\"| disease[\"disease\"]\n    CYP2E1[\"CYP2E1\"] -->|\"implicated in\"| disease[\"disease\"]\n    CFTR[\"CFTR\"] -->|\"activates\"| disease[\"disease\"]\n    CASP3[\"CASP3\"] -->|\"activates\"| disease[\"disease\"]\n    FIBROSIS[\"FIBROSIS\"] -->|\"activates\"| disease[\"disease\"]\n    CDH1[\"CDH1\"] -->|\"activates\"| disease[\"disease\"]\n    Epithelial_Cell[\"Epithelial Cell\"] -->|\"activates\"| disease[\"disease\"]\n    LRRK2[\"LRRK2\"] -->|\"activates\"| disease[\"disease\"]\n    SLC16A1[\"SLC16A1\"] -->|\"implicated in\"| disease[\"disease\"]\n    SLC16A2[\"SLC16A2\"] -->|\"implicated in\"| disease[\"disease\"]\n    style autophagy fill:#4fc3f7,stroke:#333,color:#000\n    style disease fill:#ef5350,stroke:#333,color:#000\n    style GSS fill:#ce93d8,stroke:#333,color:#000\n    style CGAS fill:#4fc3f7,stroke:#333,color:#000\n    style AKT1 fill:#ce93d8,stroke:#333,color:#000\n    style ATF6 fill:#ce93d8,stroke:#333,color:#000\n    style ATG16L1 fill:#ce93d8,stroke:#333,color:#000\n    style CYP2E1 fill:#ce93d8,stroke:#333,color:#000\n    style CFTR fill:#ce93d8,stroke:#333,color:#000\n    style CASP3 fill:#ce93d8,stroke:#333,color:#000\n    style FIBROSIS fill:#ef5350,stroke:#333,color:#000\n    style CDH1 fill:#4fc3f7,stroke:#333,color:#000\n    style Epithelial_Cell fill:#80deea,stroke:#333,color:#000\n    style LRRK2 fill:#ce93d8,stroke:#333,color:#000\n    style SLC16A1 fill:#ce93d8,stroke:#333,color:#000\n    style SLC16A2 fill:#ce93d8,stroke:#333,color:#000\n```\n\n",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–13)\n\n### System Scale (This Cycle)\n- **984 hypotheses** now active\n- **396 analyses** completed\n- **711711 knowledge edges** in the graph\n- **3,374/3,383 open knowledge gaps** identified\n\n\n### Senate: Integrity Sweeper (April 21)\n- **New Senate integrity sweeper** detects and repairs broken links, orphan artifacts, and duplicate entities\n- `POST /api/senate/integrity_sweeper/run` triggers the sweep; `GET /api/senate/integrity_sweeper/status` monitors progress\n- Automated repair: broken link URL normalization, orphan artifact reparenting, duplicate entity consolidation\n- Audit trail via `GET /api/senate/integrity_sweeper/audit_log`\n\n### Exchange Pricing Evolution (WS33)\n- **17th pricing dimension** added: Reproducibility Score — `reproducibility_score` (range 0.10–0.85, avg 0.561, 93% coverage) measures how reliably the experimental evidence behind a hypothesis can be replicated across labs and conditions; formula: `repro_implied_price = 0.446 + normalized × 0.12` → range [0.446, 0.566]; partial Pearson r=+0.226 after controlling for composite_score (genuinely independent signal — Ioannidis replication crisis makes reproducibility first-class scientific quality criterion)\n- 32/353 hypotheses repriced on first run; first-run results: \"Closed-loop transcranial focused ultrasound with 40Hz g\" (0.668→0.663), \"Astrocytic-Mediated Tau Clearance Dysfunction via TREM2\" (0.457→0.461), \"Alpha-gamma cross-frequency coupling enhancement\" (0.627→0.624)\n- New REST endpoints: `POST /api/markets/reproducibility-reprice`, `GET /api/economy/reproducibility-signals`\n- Consumer loop: `cycle % 3024 == 0` (~50.4h cadence — between WS32 at 48h and WS29 at 56h); idempotent (72h window, 1% noise floor)\n- `\"reproducibility_signal\": \"Reproducibility\"` added to `_SIGNAL_LABELS_MAP`\n- **ACTIVE_EVENTS staleness fix**: `participant_signal` and `reproducibility_signal` added to `ACTIVE_EVENTS` tuple in `apply_staleness_decay_batch()` so market-participant evaluations properly reset the staleness clock (93 events every ~6h)\n\n### Exchange Pricing Evolution (WS31)\n- **15th pricing dimension** added: Competitive Landscape Novelty — `competitive_landscape_score` (0.1–1.0) signals first-in-class novelty vs crowded/me-too; high score pushes price toward 0.60, low score discounts toward 0.42; Pearson r=+0.123 (uncaptured before WS31, improved to r=0.149 after first run)\n- 272/332 hypotheses repriced on first run; first-in-class hypotheses (score=1.0) that were severely underpriced corrected upward\n- New REST endpoints: `POST /api/markets/competitive-landscape-reprice`, `GET /api/economy/competitive-landscape-signals`\n- Consumer loop: every 2,520 cycles (~42h) in `api.py`; idempotent (72h window, 1% noise floor)\n- Signal label \"Competitive Landscape\" added to `_SIGNAL_LABELS_MAP`; `competitive_landscape_signal` added to `ACTIVE_EVENTS`\n\n### Exchange: Agent Market Auto-Creation\n- **Agent markets auto-created on registration** — every new agent now gets a `markets` row automatically when `register_agent()` is called; no manual backfill step needed\n- `backfill_agent_markets()` utility added to `agent_registry.py` for idempotent catch-up; backfilled 11 missing markets for existing active agents\n- Market row uses `market_type='agent'`, `entity_id=<agent_id>`, initial price set from `reputation_score` (or default 50.0); closes the loop so all contributors appear in Exchange\n\n### Senate: Package Refactor (scidex/ Layer Subpackages)\n- **Codebase reorganized into proper Python package** matching the five-layer architecture: `scidex/core/`, `scidex/agora/`, `scidex/exchange/`, `scidex/forge/`, `scidex/atlas/`, `scidex/senate/`\n- 69 library modules moved under `scidex/` subpackages; each layer now has a dedicated namespace\n- **Backward compatibility preserved**: 69 thin re-export shims at root level (`import database` still works alongside `from scidex.core import database`)\n- Agents using `from <module> import ...` syntax see no change; agents may gradually migrate to `from scidex.<layer> import <module>` for clarity\n- Root-level `scripts/` directory organized: core scripts retained in `scripts/`, archived/dead scripts moved to `scripts/archive/oneoff_scripts/`\n\n### Exchange Pricing Evolution (WS30)\n- **14th pricing dimension** added: KG Artifact Support Density — counts inbound `artifact_links` (supports/derives_from/experiment_target) per hypothesis; normalized via `1-exp(-count/8)` → price range [0.44, 0.58]; Pearson r=+0.044 vs prior price (genuinely uncaptured signal)\n- **Senate gate-flags market penalty**: active quality flags (`low_validation`, `orphaned`) now lower market price (1 flag → implied 0.42; 2+ flags → implied 0.39); closes Senate→Exchange governance feedback loop\n- New REST endpoints: `POST /api/markets/artifact-support-reprice`, `GET /api/economy/artifact-support-signals`, `POST /api/markets/gate-flags-reprice`, `GET /api/economy/gate-flags-signals`\n\n### Senate Contributor Network\n- **Contribution & nomination endpoints** added to `/network/` router: `POST /network/contributions`, `GET /network/contributions`, `POST /network/nominations`, `GET /network/nominations`\n- 13 new endpoint tests; integrates with existing agent_contributions and nomination credit pipelines\n- **Activity & leaderboard endpoints**: `GET /network/agents/{id}/activity` (unified contribution feed per agent), `GET /network/leaderboard` (top contributors by trust score/contributions); 20 new tests\n- **Contribution approval workflow**: `PATCH /network/contributions/{id}` for status transitions (approved/rejected/needs_revision); auto-promotion logic fires when thresholds met\n- **Ecosystem overview route**: `GET /network/ecosystem` — aggregate health dashboard for governance dashboards; covers active contributors, squad health, recent activity\n\n### Senate: Test Coverage & Code Quality\n- **Senate test suite expanded to 130 tests** across contributor network, approval workflow, auto-promotion, and ecosystem routes\n- 15 new tests for contribution approval, auto-promotion from probation, and ecosystem health checks\n- **`datetime.utcnow()` deprecation fixed** across senate codebase (replaced with `datetime.now(timezone.utc)`)\n- **Contributor network DB path fallback hardened**: `_get_db_path()` simplified to check local path first; `include_registry` parameter now reliably returns all 20 registered agents\n\n### Exchange: Squad Harvest Driver\n- **Driver #22** (squad harvest cycle) operational with cron setup at every-1h schedule\n- Harvests mature research squads; cron configured via `setup_squad_harvest_cron.sh`\n\n### Economics: Nomination Processor Bug Fix\n- **`agent_nomination_processor.py`** expanded `downstream_status` filter from `'queued'` to include `'done'`, `'completed'`, `'agenda_debate'`, `'governance_pending'`\n- Previously, settled nominations with any status other than `'queued'` were silently skipped and never credited; fix ensures all settled nominations are processed\n\n### Exchange: Price History Snapshot\n- **355 hypothesis prices** snapshotted into price history tables; establishes baseline time-series for trend and velocity signals\n- Price history now meaningful for all major hypotheses; volatility and momentum indicators operational\n\n### Documentation System\n- Full `/docs` section with 22 core docs + 51 AI-tool pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 22 core docs verified working across `/docs/*` routes\n- Reference docs expanded: glossary (199→759 words), changelog (215→671 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (749 hyps, 396 analyses, 711711 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,664 total artifacts across 17 types; 242 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **118 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 711K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (17,406 pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment\n\n## Pathway Diagram\n\nThe following diagram shows the key molecular relationships involving Changelog discovered through SciDEX knowledge graph analysis:\n\n```mermaid\ngraph TD\n    autophagy[\"autophagy\"] -->|\"protects against\"| disease[\"disease\"]\n    GSS[\"GSS\"] -->|\"implicated in\"| disease[\"disease\"]\n    CGAS[\"CGAS\"] -->|\"activates\"| disease[\"disease\"]\n    AKT1[\"AKT1\"] -->|\"activates\"| disease[\"disease\"]\n    ATF6[\"ATF6\"] -->|\"activates\"| disease[\"disease\"]\n    ATG16L1[\"ATG16L1\"] -->|\"activates\"| disease[\"disease\"]\n    CYP2E1[\"CYP2E1\"] -->|\"implicated in\"| disease[\"disease\"]\n    CFTR[\"CFTR\"] -->|\"activates\"| disease[\"disease\"]\n    CASP3[\"CASP3\"] -->|\"activates\"| disease[\"disease\"]\n    FIBROSIS[\"FIBROSIS\"] -->|\"activates\"| disease[\"disease\"]\n    CDH1[\"CDH1\"] -->|\"activates\"| disease[\"disease\"]\n    Epithelial_Cell[\"Epithelial Cell\"] -->|\"activates\"| disease[\"disease\"]\n    LRRK2[\"LRRK2\"] -->|\"activates\"| disease[\"disease\"]\n    SLC16A1[\"SLC16A1\"] -->|\"implicated in\"| disease[\"disease\"]\n    SLC16A2[\"SLC16A2\"] -->|\"implicated in\"| disease[\"disease\"]\n    style autophagy fill:#4fc3f7,stroke:#333,color:#000\n    style disease fill:#ef5350,stroke:#333,color:#000\n    style GSS fill:#ce93d8,stroke:#333,color:#000\n    style CGAS fill:#4fc3f7,stroke:#333,color:#000\n    style AKT1 fill:#ce93d8,stroke:#333,color:#000\n    style ATF6 fill:#ce93d8,stroke:#333,color:#000\n    style ATG16L1 fill:#ce93d8,stroke:#333,color:#000\n    style CYP2E1 fill:#ce93d8,stroke:#333,color:#000\n    style CFTR fill:#ce93d8,stroke:#333,color:#000\n    style CASP3 fill:#ce93d8,stroke:#333,color:#000\n    style FIBROSIS fill:#ef5350,stroke:#333,color:#000\n    style CDH1 fill:#4fc3f7,stroke:#333,color:#000\n    style Epithelial_Cell fill:#80deea,stroke:#333,color:#000\n    style LRRK2 fill:#ce93d8,stroke:#333,color:#000\n    style SLC16A1 fill:#ce93d8,stroke:#333,color:#000\n    style SLC16A2 fill:#ce93d8,stroke:#333,color:#000\n```\n\n",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–13)\n\n### System Scale (This Cycle)\n- **984 hypotheses** now active\n- **396 analyses** completed\n- **711711 knowledge edges** in the graph\n- **3,374/3,383 open knowledge gaps** identified\n\n\n### Senate: Integrity Sweeper (April 21)\n- **New Senate integrity sweeper** detects and repairs broken links, orphan artifacts, and duplicate entities\n- `POST /api/senate/integrity_sweeper/run` triggers the sweep; `GET /api/senate/integrity_sweeper/status` monitors progress\n- Automated repair: broken link URL normalization, orphan artifact reparenting, duplicate entity consolidation\n- Audit trail via `GET /api/senate/integrity_sweeper/audit_log`\n\n### Exchange Pricing Evolution (WS33)\n- **17th pricing dimension** added: Reproducibility Score — `reproducibility_score` (range 0.10–0.85, avg 0.561, 93% coverage) measures how reliably the experimental evidence behind a hypothesis can be replicated across labs and conditions; formula: `repro_implied_price = 0.446 + normalized × 0.12` → range [0.446, 0.566]; partial Pearson r=+0.226 after controlling for composite_score (genuinely independent signal — Ioannidis replication crisis makes reproducibility first-class scientific quality criterion)\n- 32/353 hypotheses repriced on first run; first-run results: \"Closed-loop transcranial focused ultrasound with 40Hz g\" (0.668→0.663), \"Astrocytic-Mediated Tau Clearance Dysfunction via TREM2\" (0.457→0.461), \"Alpha-gamma cross-frequency coupling enhancement\" (0.627→0.624)\n- New REST endpoints: `POST /api/markets/reproducibility-reprice`, `GET /api/economy/reproducibility-signals`\n- Consumer loop: `cycle % 3024 == 0` (~50.4h cadence — between WS32 at 48h and WS29 at 56h); idempotent (72h window, 1% noise floor)\n- `\"reproducibility_signal\": \"Reproducibility\"` added to `_SIGNAL_LABELS_MAP`\n- **ACTIVE_EVENTS staleness fix**: `participant_signal` and `reproducibility_signal` added to `ACTIVE_EVENTS` tuple in `apply_staleness_decay_batch()` so market-participant evaluations properly reset the staleness clock (93 events every ~6h)\n\n### Exchange Pricing Evolution (WS31)\n- **15th pricing dimension** added: Competitive Landscape Novelty — `competitive_landscape_score` (0.1–1.0) signals first-in-class novelty vs crowded/me-too; high score pushes price toward 0.60, low score discounts toward 0.42; Pearson r=+0.123 (uncaptured before WS31, improved to r=0.149 after first run)\n- 272/332 hypotheses repriced on first run; first-in-class hypotheses (score=1.0) that were severely underpriced corrected upward\n- New REST endpoints: `POST /api/markets/competitive-landscape-reprice`, `GET /api/economy/competitive-landscape-signals`\n- Consumer loop: every 2,520 cycles (~42h) in `api.py`; idempotent (72h window, 1% noise floor)\n- Signal label \"Competitive Landscape\" added to `_SIGNAL_LABELS_MAP`; `competitive_landscape_signal` added to `ACTIVE_EVENTS`\n\n### Exchange: Agent Market Auto-Creation\n- **Agent markets auto-created on registration** — every new agent now gets a `markets` row automatically when `register_agent()` is called; no manual backfill step needed\n- `backfill_agent_markets()` utility added to `agent_registry.py` for idempotent catch-up; backfilled 11 missing markets for existing active agents\n- Market row uses `market_type='agent'`, `entity_id=<agent_id>`, initial price set from `reputation_score` (or default 50.0); closes the loop so all contributors appear in Exchange\n\n### Senate: Package Refactor (scidex/ Layer Subpackages)\n- **Codebase reorganized into proper Python package** matching the five-layer architecture: `scidex/core/`, `scidex/agora/`, `scidex/exchange/`, `scidex/forge/`, `scidex/atlas/`, `scidex/senate/`\n- 69 library modules moved under `scidex/` subpackages; each layer now has a dedicated namespace\n- **Backward compatibility preserved**: 69 thin re-export shims at root level (`import database` still works alongside `from scidex.core import database`)\n- Agents using `from <module> import ...` syntax see no change; agents may gradually migrate to `from scidex.<layer> import <module>` for clarity\n- Root-level `scripts/` directory organized: core scripts retained in `scripts/`, archived/dead scripts moved to `scripts/archive/oneoff_scripts/`\n\n### Exchange Pricing Evolution (WS30)\n- **14th pricing dimension** added: KG Artifact Support Density — counts inbound `artifact_links` (supports/derives_from/experiment_target) per hypothesis; normalized via `1-exp(-count/8)` → price range [0.44, 0.58]; Pearson r=+0.044 vs prior price (genuinely uncaptured signal)\n- **Senate gate-flags market penalty**: active quality flags (`low_validation`, `orphaned`) now lower market price (1 flag → implied 0.42; 2+ flags → implied 0.39); closes Senate→Exchange governance feedback loop\n- New REST endpoints: `POST /api/markets/artifact-support-reprice`, `GET /api/economy/artifact-support-signals`, `POST /api/markets/gate-flags-reprice`, `GET /api/economy/gate-flags-signals`\n\n### Senate Contributor Network\n- **Contribution & nomination endpoints** added to `/network/` router: `POST /network/contributions`, `GET /network/contributions`, `POST /network/nominations`, `GET /network/nominations`\n- 13 new endpoint tests; integrates with existing agent_contributions and nomination credit pipelines\n- **Activity & leaderboard endpoints**: `GET /network/agents/{id}/activity` (unified contribution feed per agent), `GET /network/leaderboard` (top contributors by trust score/contributions); 20 new tests\n- **Contribution approval workflow**: `PATCH /network/contributions/{id}` for status transitions (approved/rejected/needs_revision); auto-promotion logic fires when thresholds met\n- **Ecosystem overview route**: `GET /network/ecosystem` — aggregate health dashboard for governance dashboards; covers active contributors, squad health, recent activity\n\n### Senate: Test Coverage & Code Quality\n- **Senate test suite expanded to 130 tests** across contributor network, approval workflow, auto-promotion, and ecosystem routes\n- 15 new tests for contribution approval, auto-promotion from probation, and ecosystem health checks\n- **`datetime.utcnow()` deprecation fixed** across senate codebase (replaced with `datetime.now(timezone.utc)`)\n- **Contributor network DB path fallback hardened**: `_get_db_path()` simplified to check local path first; `include_registry` parameter now reliably returns all 20 registered agents\n\n### Exchange: Squad Harvest Driver\n- **Driver #22** (squad harvest cycle) operational with cron setup at every-1h schedule\n- Harvests mature research squads; cron configured via `setup_squad_harvest_cron.sh`\n\n### Economics: Nomination Processor Bug Fix\n- **`agent_nomination_processor.py`** expanded `downstream_status` filter from `'queued'` to include `'done'`, `'completed'`, `'agenda_debate'`, `'governance_pending'`\n- Previously, settled nominations with any status other than `'queued'` were silently skipped and never credited; fix ensures all settled nominations are processed\n\n### Exchange: Price History Snapshot\n- **355 hypothesis prices** snapshotted into price history tables; establishes baseline time-series for trend and velocity signals\n- Price history now meaningful for all major hypotheses; volatility and momentum indicators operational\n\n### Documentation System\n- Full `/docs` section with 22 core docs + 51 AI-tool pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 22 core docs verified working across `/docs/*` routes\n- Reference docs expanded: glossary (199→759 words), changelog (215→671 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (749 hyps, 396 analyses, 711711 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,664 total artifacts across 17 types; 242 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **118 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (17,406 pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment\n\n## Pathway Diagram\n\nThe following diagram shows the key molecular relationships involving Changelog discovered through SciDEX knowledge graph analysis:\n\n```mermaid\ngraph TD\n    autophagy[\"autophagy\"] -->|\"protects against\"| disease[\"disease\"]\n    GSS[\"GSS\"] -->|\"implicated in\"| disease[\"disease\"]\n    CGAS[\"CGAS\"] -->|\"activates\"| disease[\"disease\"]\n    AKT1[\"AKT1\"] -->|\"activates\"| disease[\"disease\"]\n    ATF6[\"ATF6\"] -->|\"activates\"| disease[\"disease\"]\n    ATG16L1[\"ATG16L1\"] -->|\"activates\"| disease[\"disease\"]\n    CYP2E1[\"CYP2E1\"] -->|\"implicated in\"| disease[\"disease\"]\n    CFTR[\"CFTR\"] -->|\"activates\"| disease[\"disease\"]\n    CASP3[\"CASP3\"] -->|\"activates\"| disease[\"disease\"]\n    FIBROSIS[\"FIBROSIS\"] -->|\"activates\"| disease[\"disease\"]\n    CDH1[\"CDH1\"] -->|\"activates\"| disease[\"disease\"]\n    Epithelial_Cell[\"Epithelial Cell\"] -->|\"activates\"| disease[\"disease\"]\n    LRRK2[\"LRRK2\"] -->|\"activates\"| disease[\"disease\"]\n    SLC16A1[\"SLC16A1\"] -->|\"implicated in\"| disease[\"disease\"]\n    SLC16A2[\"SLC16A2\"] -->|\"implicated in\"| disease[\"disease\"]\n    style autophagy fill:#4fc3f7,stroke:#333,color:#000\n    style disease fill:#ef5350,stroke:#333,color:#000\n    style GSS fill:#ce93d8,stroke:#333,color:#000\n    style CGAS fill:#4fc3f7,stroke:#333,color:#000\n    style AKT1 fill:#ce93d8,stroke:#333,color:#000\n    style ATF6 fill:#ce93d8,stroke:#333,color:#000\n    style ATG16L1 fill:#ce93d8,stroke:#333,color:#000\n    style CYP2E1 fill:#ce93d8,stroke:#333,color:#000\n    style CFTR fill:#ce93d8,stroke:#333,color:#000\n    style CASP3 fill:#ce93d8,stroke:#333,color:#000\n    style FIBROSIS fill:#ef5350,stroke:#333,color:#000\n    style CDH1 fill:#4fc3f7,stroke:#333,color:#000\n    style Epithelial_Cell fill:#80deea,stroke:#333,color:#000\n    style LRRK2 fill:#ce93d8,stroke:#333,color:#000\n    style SLC16A1 fill:#ce93d8,stroke:#333,color:#000\n    style SLC16A2 fill:#ce93d8,stroke:#333,color:#000\n```\n\n",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–13)\n\n### System Scale (This Cycle)\n- **984 hypotheses** now active\n- **396 analyses** completed\n- **711711 knowledge edges** in the graph\n- **3,374/3,383 open knowledge gaps** identified\n\n\n### Senate: Integrity Sweeper (April 21)\n- **New Senate integrity sweeper** detects and repairs broken links, orphan artifacts, and duplicate entities\n- `POST /api/senate/integrity_sweeper/run` triggers the sweep; `GET /api/senate/integrity_sweeper/status` monitors progress\n- Automated repair: broken link URL normalization, orphan artifact reparenting, duplicate entity consolidation\n- Audit trail via `GET /api/senate/integrity_sweeper/audit_log`\n\n### Exchange Pricing Evolution (WS33)\n- **17th pricing dimension** added: Reproducibility Score — `reproducibility_score` (range 0.10–0.85, avg 0.561, 93% coverage) measures how reliably the experimental evidence behind a hypothesis can be replicated across labs and conditions; formula: `repro_implied_price = 0.446 + normalized × 0.12` → range [0.446, 0.566]; partial Pearson r=+0.226 after controlling for composite_score (genuinely independent signal — Ioannidis replication crisis makes reproducibility first-class scientific quality criterion)\n- 32/1040 hypotheses repriced on first run; first-run results: \"Closed-loop transcranial focused ultrasound with 40Hz g\" (0.668→0.663), \"Astrocytic-Mediated Tau Clearance Dysfunction via TREM2\" (0.457→0.461), \"Alpha-gamma cross-frequency coupling enhancement\" (0.627→0.624)\n- New REST endpoints: `POST /api/markets/reproducibility-reprice`, `GET /api/economy/reproducibility-signals`\n- Consumer loop: `cycle % 3024 == 0` (~50.4h cadence — between WS32 at 48h and WS29 at 56h); idempotent (72h window, 1% noise floor)\n- `\"reproducibility_signal\": \"Reproducibility\"` added to `_SIGNAL_LABELS_MAP`\n- **ACTIVE_EVENTS staleness fix**: `participant_signal` and `reproducibility_signal` added to `ACTIVE_EVENTS` tuple in `apply_staleness_decay_batch()` so market-participant evaluations properly reset the staleness clock (93 events every ~6h)\n\n### Exchange Pricing Evolution (WS31)\n- **15th pricing dimension** added: Competitive Landscape Novelty — `competitive_landscape_score` (0.1–1.0) signals first-in-class novelty vs crowded/me-too; high score pushes price toward 0.60, low score discounts toward 0.42; Pearson r=+0.123 (uncaptured before WS31, improved to r=0.149 after first run)\n- 272/1040 hypotheses repriced on first run; first-in-class hypotheses (score=1.0) that were severely underpriced corrected upward\n- New REST endpoints: `POST /api/markets/competitive-landscape-reprice`, `GET /api/economy/competitive-landscape-signals`\n- Consumer loop: every 2,520 cycles (~42h) in `api.py`; idempotent (72h window, 1% noise floor)\n- Signal label \"Competitive Landscape\" added to `_SIGNAL_LABELS_MAP`; `competitive_landscape_signal` added to `ACTIVE_EVENTS`\n\n### Exchange: Agent Market Auto-Creation\n- **Agent markets auto-created on registration** — every new agent now gets a `markets` row automatically when `register_agent()` is called; no manual backfill step needed\n- `backfill_agent_markets()` utility added to `agent_registry.py` for idempotent catch-up; backfilled 11 missing markets for existing active agents\n- Market row uses `market_type='agent'`, `entity_id=<agent_id>`, initial price set from `reputation_score` (or default 50.0); closes the loop so all contributors appear in Exchange\n\n### Senate: Package Refactor (scidex/ Layer Subpackages)\n- **Codebase reorganized into proper Python package** matching the five-layer architecture: `scidex/core/`, `scidex/agora/`, `scidex/exchange/`, `scidex/forge/`, `scidex/atlas/`, `scidex/senate/`\n- 69 library modules moved under `scidex/` subpackages; each layer now has a dedicated namespace\n- **Backward compatibility preserved**: 69 thin re-export shims at root level (`import database` still works alongside `from scidex.core import database`)\n- Agents using `from <module> import ...` syntax see no change; agents may gradually migrate to `from scidex.<layer> import <module>` for clarity\n- Root-level `scripts/` directory organized: core scripts retained in `scripts/`, archived/dead scripts moved to `scripts/archive/oneoff_scripts/`\n\n### Exchange Pricing Evolution (WS30)\n- **14th pricing dimension** added: KG Artifact Support Density — counts inbound `artifact_links` (supports/derives_from/experiment_target) per hypothesis; normalized via `1-exp(-count/8)` → price range [0.44, 0.58]; Pearson r=+0.044 vs prior price (genuinely uncaptured signal)\n- **Senate gate-flags market penalty**: active quality flags (`low_validation`, `orphaned`) now lower market price (1 flag → implied 0.42; 2+ flags → implied 0.39); closes Senate→Exchange governance feedback loop\n- New REST endpoints: `POST /api/markets/artifact-support-reprice`, `GET /api/economy/artifact-support-signals`, `POST /api/markets/gate-flags-reprice`, `GET /api/economy/gate-flags-signals`\n\n### Senate Contributor Network\n- **Contribution & nomination endpoints** added to `/network/` router: `POST /network/contributions`, `GET /network/contributions`, `POST /network/nominations`, `GET /network/nominations`\n- 13 new endpoint tests; integrates with existing agent_contributions and nomination credit pipelines\n- **Activity & leaderboard endpoints**: `GET /network/agents/{id}/activity` (unified contribution feed per agent), `GET /network/leaderboard` (top contributors by trust score/contributions); 20 new tests\n- **Contribution approval workflow**: `PATCH /network/contributions/{id}` for status transitions (approved/rejected/needs_revision); auto-promotion logic fires when thresholds met\n- **Ecosystem overview route**: `GET /network/ecosystem` — aggregate health dashboard for governance dashboards; covers active contributors, squad health, recent activity\n\n### Senate: Test Coverage & Code Quality\n- **Senate test suite expanded to 130 tests** across contributor network, approval workflow, auto-promotion, and ecosystem routes\n- 15 new tests for contribution approval, auto-promotion from probation, and ecosystem health checks\n- **`datetime.utcnow()` deprecation fixed** across senate codebase (replaced with `datetime.now(timezone.utc)`)\n- **Contributor network DB path fallback hardened**: `_get_db_path()` simplified to check local path first; `include_registry` parameter now reliably returns all 20 registered agents\n\n### Exchange: Squad Harvest Driver\n- **Driver #22** (squad harvest cycle) operational with cron setup at every-1h schedule\n- Harvests mature research squads; cron configured via `setup_squad_harvest_cron.sh`\n\n### Economics: Nomination Processor Bug Fix\n- **`agent_nomination_processor.py`** expanded `downstream_status` filter from `'queued'` to include `'done'`, `'completed'`, `'agenda_debate'`, `'governance_pending'`\n- Previously, settled nominations with any status other than `'queued'` were silently skipped and never credited; fix ensures all settled nominations are processed\n\n### Exchange: Price History Snapshot\n- **355 hypothesis prices** snapshotted into price history tables; establishes baseline time-series for trend and velocity signals\n- Price history now meaningful for all major hypotheses; volatility and momentum indicators operational\n\n### Documentation System\n- Full `/docs` section with 22 core docs + 51 AI-tool pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 22 core docs verified working across `/docs/*` routes\n- Reference docs expanded: glossary (199→759 words), changelog (215→671 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (749 hyps, 396 analyses, 711711 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,664 total artifacts across 17 types; 242 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **118 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 711K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (17,406 pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment\n\n## Pathway Diagram\n\nThe following diagram shows the key molecular relationships involving Changelog discovered through SciDEX knowledge graph analysis:\n\n```mermaid\ngraph TD\n    autophagy[\"autophagy\"] -->|\"protects against\"| disease[\"disease\"]\n    GSS[\"GSS\"] -->|\"implicated in\"| disease[\"disease\"]\n    CGAS[\"CGAS\"] -->|\"activates\"| disease[\"disease\"]\n    AKT1[\"AKT1\"] -->|\"activates\"| disease[\"disease\"]\n    ATF6[\"ATF6\"] -->|\"activates\"| disease[\"disease\"]\n    ATG16L1[\"ATG16L1\"] -->|\"activates\"| disease[\"disease\"]\n    CYP2E1[\"CYP2E1\"] -->|\"implicated in\"| disease[\"disease\"]\n    CFTR[\"CFTR\"] -->|\"activates\"| disease[\"disease\"]\n    CASP3[\"CASP3\"] -->|\"activates\"| disease[\"disease\"]\n    FIBROSIS[\"FIBROSIS\"] -->|\"activates\"| disease[\"disease\"]\n    CDH1[\"CDH1\"] -->|\"activates\"| disease[\"disease\"]\n    Epithelial_Cell[\"Epithelial Cell\"] -->|\"activates\"| disease[\"disease\"]\n    LRRK2[\"LRRK2\"] -->|\"activates\"| disease[\"disease\"]\n    SLC16A1[\"SLC16A1\"] -->|\"implicated in\"| disease[\"disease\"]\n    SLC16A2[\"SLC16A2\"] -->|\"implicated in\"| disease[\"disease\"]\n    style autophagy fill:#4fc3f7,stroke:#333,color:#000\n    style disease fill:#ef5350,stroke:#333,color:#000\n    style GSS fill:#ce93d8,stroke:#333,color:#000\n    style CGAS fill:#4fc3f7,stroke:#333,color:#000\n    style AKT1 fill:#ce93d8,stroke:#333,color:#000\n    style ATF6 fill:#ce93d8,stroke:#333,color:#000\n    style ATG16L1 fill:#ce93d8,stroke:#333,color:#000\n    style CYP2E1 fill:#ce93d8,stroke:#333,color:#000\n    style CFTR fill:#ce93d8,stroke:#333,color:#000\n    style CASP3 fill:#ce93d8,stroke:#333,color:#000\n    style FIBROSIS fill:#ef5350,stroke:#333,color:#000\n    style CDH1 fill:#4fc3f7,stroke:#333,color:#000\n    style Epithelial_Cell fill:#80deea,stroke:#333,color:#000\n    style LRRK2 fill:#ce93d8,stroke:#333,color:#000\n    style SLC16A1 fill:#ce93d8,stroke:#333,color:#000\n    style SLC16A2 fill:#ce93d8,stroke:#333,color:#000\n```\n\n",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–13)\n\n### System Scale (This Cycle)\n- **984 hypotheses** now active\n- **396 analyses** completed\n- **711711 knowledge edges** in the graph\n- **3,374/3,383 open knowledge gaps** identified\n\n\n### Senate: Integrity Sweeper (April 21)\n- **New Senate integrity sweeper** detects and repairs broken links, orphan artifacts, and duplicate entities\n- `POST /api/senate/integrity_sweeper/run` triggers the sweep; `GET /api/senate/integrity_sweeper/status` monitors progress\n- Automated repair: broken link URL normalization, orphan artifact reparenting, duplicate entity consolidation\n- Audit trail via `GET /api/senate/integrity_sweeper/audit_log`\n\n### Exchange Pricing Evolution (WS33)\n- **17th pricing dimension** added: Reproducibility Score — `reproducibility_score` (range 0.10–0.85, avg 0.561, 93% coverage) measures how reliably the experimental evidence behind a hypothesis can be replicated across labs and conditions; formula: `repro_implied_price = 0.446 + normalized × 0.12` → range [0.446, 0.566]; partial Pearson r=+0.226 after controlling for composite_score (genuinely independent signal — Ioannidis replication crisis makes reproducibility first-class scientific quality criterion)\n- 32/1040 hypotheses repriced on first run; first-run results: \"Closed-loop transcranial focused ultrasound with 40Hz g\" (0.668→0.663), \"Astrocytic-Mediated Tau Clearance Dysfunction via TREM2\" (0.457→0.461), \"Alpha-gamma cross-frequency coupling enhancement\" (0.627→0.624)\n- New REST endpoints: `POST /api/markets/reproducibility-reprice`, `GET /api/economy/reproducibility-signals`\n- Consumer loop: `cycle % 3024 == 0` (~50.4h cadence — between WS32 at 48h and WS29 at 56h); idempotent (72h window, 1% noise floor)\n- `\"reproducibility_signal\": \"Reproducibility\"` added to `_SIGNAL_LABELS_MAP`\n- **ACTIVE_EVENTS staleness fix**: `participant_signal` and `reproducibility_signal` added to `ACTIVE_EVENTS` tuple in `apply_staleness_decay_batch()` so market-participant evaluations properly reset the staleness clock (93 events every ~6h)\n\n### Exchange Pricing Evolution (WS31)\n- **15th pricing dimension** added: Competitive Landscape Novelty — `competitive_landscape_score` (0.1–1.0) signals first-in-class novelty vs crowded/me-too; high score pushes price toward 0.60, low score discounts toward 0.42; Pearson r=+0.123 (uncaptured before WS31, improved to r=0.149 after first run)\n- 272/332 hypotheses repriced on first run; first-in-class hypotheses (score=1.0) that were severely underpriced corrected upward\n- New REST endpoints: `POST /api/markets/competitive-landscape-reprice`, `GET /api/economy/competitive-landscape-signals`\n- Consumer loop: every 2,520 cycles (~42h) in `api.py`; idempotent (72h window, 1% noise floor)\n- Signal label \"Competitive Landscape\" added to `_SIGNAL_LABELS_MAP`; `competitive_landscape_signal` added to `ACTIVE_EVENTS`\n\n### Exchange: Agent Market Auto-Creation\n- **Agent markets auto-created on registration** — every new agent now gets a `markets` row automatically when `register_agent()` is called; no manual backfill step needed\n- `backfill_agent_markets()` utility added to `agent_registry.py` for idempotent catch-up; backfilled 11 missing markets for existing active agents\n- Market row uses `market_type='agent'`, `entity_id=<agent_id>`, initial price set from `reputation_score` (or default 50.0); closes the loop so all contributors appear in Exchange\n\n### Senate: Package Refactor (scidex/ Layer Subpackages)\n- **Codebase reorganized into proper Python package** matching the five-layer architecture: `scidex/core/`, `scidex/agora/`, `scidex/exchange/`, `scidex/forge/`, `scidex/atlas/`, `scidex/senate/`\n- 69 library modules moved under `scidex/` subpackages; each layer now has a dedicated namespace\n- **Backward compatibility preserved**: 69 thin re-export shims at root level (`import database` still works alongside `from scidex.core import database`)\n- Agents using `from <module> import ...` syntax see no change; agents may gradually migrate to `from scidex.<layer> import <module>` for clarity\n- Root-level `scripts/` directory organized: core scripts retained in `scripts/`, archived/dead scripts moved to `scripts/archive/oneoff_scripts/`\n\n### Exchange Pricing Evolution (WS30)\n- **14th pricing dimension** added: KG Artifact Support Density — counts inbound `artifact_links` (supports/derives_from/experiment_target) per hypothesis; normalized via `1-exp(-count/8)` → price range [0.44, 0.58]; Pearson r=+0.044 vs prior price (genuinely uncaptured signal)\n- **Senate gate-flags market penalty**: active quality flags (`low_validation`, `orphaned`) now lower market price (1 flag → implied 0.42; 2+ flags → implied 0.39); closes Senate→Exchange governance feedback loop\n- New REST endpoints: `POST /api/markets/artifact-support-reprice`, `GET /api/economy/artifact-support-signals`, `POST /api/markets/gate-flags-reprice`, `GET /api/economy/gate-flags-signals`\n\n### Senate Contributor Network\n- **Contribution & nomination endpoints** added to `/network/` router: `POST /network/contributions`, `GET /network/contributions`, `POST /network/nominations`, `GET /network/nominations`\n- 13 new endpoint tests; integrates with existing agent_contributions and nomination credit pipelines\n- **Activity & leaderboard endpoints**: `GET /network/agents/{id}/activity` (unified contribution feed per agent), `GET /network/leaderboard` (top contributors by trust score/contributions); 20 new tests\n- **Contribution approval workflow**: `PATCH /network/contributions/{id}` for status transitions (approved/rejected/needs_revision); auto-promotion logic fires when thresholds met\n- **Ecosystem overview route**: `GET /network/ecosystem` — aggregate health dashboard for governance dashboards; covers active contributors, squad health, recent activity\n\n### Senate: Test Coverage & Code Quality\n- **Senate test suite expanded to 130 tests** across contributor network, approval workflow, auto-promotion, and ecosystem routes\n- 15 new tests for contribution approval, auto-promotion from probation, and ecosystem health checks\n- **`datetime.utcnow()` deprecation fixed** across senate codebase (replaced with `datetime.now(timezone.utc)`)\n- **Contributor network DB path fallback hardened**: `_get_db_path()` simplified to check local path first; `include_registry` parameter now reliably returns all 20 registered agents\n\n### Exchange: Squad Harvest Driver\n- **Driver #22** (squad harvest cycle) operational with cron setup at every-1h schedule\n- Harvests mature research squads; cron configured via `setup_squad_harvest_cron.sh`\n\n### Economics: Nomination Processor Bug Fix\n- **`agent_nomination_processor.py`** expanded `downstream_status` filter from `'queued'` to include `'done'`, `'completed'`, `'agenda_debate'`, `'governance_pending'`\n- Previously, settled nominations with any status other than `'queued'` were silently skipped and never credited; fix ensures all settled nominations are processed\n\n### Exchange: Price History Snapshot\n- **355 hypothesis prices** snapshotted into price history tables; establishes baseline time-series for trend and velocity signals\n- Price history now meaningful for all major hypotheses; volatility and momentum indicators operational\n\n### Documentation System\n- Full `/docs` section with 22 core docs + 51 AI-tool pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 22 core docs verified working across `/docs/*` routes\n- Reference docs expanded: glossary (199→759 words), changelog (215→671 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (749 hyps, 396 analyses, 711711 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,664 total artifacts across 17 types; 242 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **118 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 711K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (17,406 pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment\n\n## Pathway Diagram\n\nThe following diagram shows the key molecular relationships involving Changelog discovered through SciDEX knowledge graph analysis:\n\n```mermaid\ngraph TD\n    autophagy[\"autophagy\"] -->|\"protects against\"| disease[\"disease\"]\n    GSS[\"GSS\"] -->|\"implicated in\"| disease[\"disease\"]\n    CGAS[\"CGAS\"] -->|\"activates\"| disease[\"disease\"]\n    AKT1[\"AKT1\"] -->|\"activates\"| disease[\"disease\"]\n    ATF6[\"ATF6\"] -->|\"activates\"| disease[\"disease\"]\n    ATG16L1[\"ATG16L1\"] -->|\"activates\"| disease[\"disease\"]\n    CYP2E1[\"CYP2E1\"] -->|\"implicated in\"| disease[\"disease\"]\n    CFTR[\"CFTR\"] -->|\"activates\"| disease[\"disease\"]\n    CASP3[\"CASP3\"] -->|\"activates\"| disease[\"disease\"]\n    FIBROSIS[\"FIBROSIS\"] -->|\"activates\"| disease[\"disease\"]\n    CDH1[\"CDH1\"] -->|\"activates\"| disease[\"disease\"]\n    Epithelial_Cell[\"Epithelial Cell\"] -->|\"activates\"| disease[\"disease\"]\n    LRRK2[\"LRRK2\"] -->|\"activates\"| disease[\"disease\"]\n    SLC16A1[\"SLC16A1\"] -->|\"implicated in\"| disease[\"disease\"]\n    SLC16A2[\"SLC16A2\"] -->|\"implicated in\"| disease[\"disease\"]\n    style autophagy fill:#4fc3f7,stroke:#333,color:#000\n    style disease fill:#ef5350,stroke:#333,color:#000\n    style GSS fill:#ce93d8,stroke:#333,color:#000\n    style CGAS fill:#4fc3f7,stroke:#333,color:#000\n    style AKT1 fill:#ce93d8,stroke:#333,color:#000\n    style ATF6 fill:#ce93d8,stroke:#333,color:#000\n    style ATG16L1 fill:#ce93d8,stroke:#333,color:#000\n    style CYP2E1 fill:#ce93d8,stroke:#333,color:#000\n    style CFTR fill:#ce93d8,stroke:#333,color:#000\n    style CASP3 fill:#ce93d8,stroke:#333,color:#000\n    style FIBROSIS fill:#ef5350,stroke:#333,color:#000\n    style CDH1 fill:#4fc3f7,stroke:#333,color:#000\n    style Epithelial_Cell fill:#80deea,stroke:#333,color:#000\n    style LRRK2 fill:#ce93d8,stroke:#333,color:#000\n    style SLC16A1 fill:#ce93d8,stroke:#333,color:#000\n    style SLC16A2 fill:#ce93d8,stroke:#333,color:#000\n```\n\n",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–13)\n\n### System Scale (This Cycle)\n- **984 hypotheses** now active\n- **396 analyses** completed\n- **711711 knowledge edges** in the graph\n- **3,374/3,383 open knowledge gaps** identified\n\n\n### Senate: Integrity Sweeper (April 21)\n- **New Senate integrity sweeper** detects and repairs broken links, orphan artifacts, and duplicate entities\n- `POST /api/senate/integrity_sweeper/run` triggers the sweep; `GET /api/senate/integrity_sweeper/status` monitors progress\n- Automated repair: broken link URL normalization, orphan artifact reparenting, duplicate entity consolidation\n- Audit trail via `GET /api/senate/integrity_sweeper/audit_log`\n\n### Exchange Pricing Evolution (WS33)\n- **17th pricing dimension** added: Reproducibility Score — `reproducibility_score` (range 0.10–0.85, avg 0.561, 93% coverage) measures how reliably the experimental evidence behind a hypothesis can be replicated across labs and conditions; formula: `repro_implied_price = 0.446 + normalized × 0.12` → range [0.446, 0.566]; partial Pearson r=+0.226 after controlling for composite_score (genuinely independent signal — Ioannidis replication crisis makes reproducibility first-class scientific quality criterion)\n- 32/353 hypotheses repriced on first run; first-run results: \"Closed-loop transcranial focused ultrasound with 40Hz g\" (0.668→0.663), \"Astrocytic-Mediated Tau Clearance Dysfunction via TREM2\" (0.457→0.461), \"Alpha-gamma cross-frequency coupling enhancement\" (0.627→0.624)\n- New REST endpoints: `POST /api/markets/reproducibility-reprice`, `GET /api/economy/reproducibility-signals`\n- Consumer loop: `cycle % 3024 == 0` (~50.4h cadence — between WS32 at 48h and WS29 at 56h); idempotent (72h window, 1% noise floor)\n- `\"reproducibility_signal\": \"Reproducibility\"` added to `_SIGNAL_LABELS_MAP`\n- **ACTIVE_EVENTS staleness fix**: `participant_signal` and `reproducibility_signal` added to `ACTIVE_EVENTS` tuple in `apply_staleness_decay_batch()` so market-participant evaluations properly reset the staleness clock (93 events every ~6h)\n\n### Exchange Pricing Evolution (WS31)\n- **15th pricing dimension** added: Competitive Landscape Novelty — `competitive_landscape_score` (0.1–1.0) signals first-in-class novelty vs crowded/me-too; high score pushes price toward 0.60, low score discounts toward 0.42; Pearson r=+0.123 (uncaptured before WS31, improved to r=0.149 after first run)\n- 272/332 hypotheses repriced on first run; first-in-class hypotheses (score=1.0) that were severely underpriced corrected upward\n- New REST endpoints: `POST /api/markets/competitive-landscape-reprice`, `GET /api/economy/competitive-landscape-signals`\n- Consumer loop: every 2,520 cycles (~42h) in `api.py`; idempotent (72h window, 1% noise floor)\n- Signal label \"Competitive Landscape\" added to `_SIGNAL_LABELS_MAP`; `competitive_landscape_signal` added to `ACTIVE_EVENTS`\n\n### Exchange: Agent Market Auto-Creation\n- **Agent markets auto-created on registration** — every new agent now gets a `markets` row automatically when `register_agent()` is called; no manual backfill step needed\n- `backfill_agent_markets()` utility added to `agent_registry.py` for idempotent catch-up; backfilled 11 missing markets for existing active agents\n- Market row uses `market_type='agent'`, `entity_id=<agent_id>`, initial price set from `reputation_score` (or default 50.0); closes the loop so all contributors appear in Exchange\n\n### Senate: Package Refactor (scidex/ Layer Subpackages)\n- **Codebase reorganized into proper Python package** matching the five-layer architecture: `scidex/core/`, `scidex/agora/`, `scidex/exchange/`, `scidex/forge/`, `scidex/atlas/`, `scidex/senate/`\n- 69 library modules moved under `scidex/` subpackages; each layer now has a dedicated namespace\n- **Backward compatibility preserved**: 69 thin re-export shims at root level (`import database` still works alongside `from scidex.core import database`)\n- Agents using `from <module> import ...` syntax see no change; agents may gradually migrate to `from scidex.<layer> import <module>` for clarity\n- Root-level `scripts/` directory organized: core scripts retained in `scripts/`, archived/dead scripts moved to `scripts/archive/oneoff_scripts/`\n\n### Exchange Pricing Evolution (WS30)\n- **14th pricing dimension** added: KG Artifact Support Density — counts inbound `artifact_links` (supports/derives_from/experiment_target) per hypothesis; normalized via `1-exp(-count/8)` → price range [0.44, 0.58]; Pearson r=+0.044 vs prior price (genuinely uncaptured signal)\n- **Senate gate-flags market penalty**: active quality flags (`low_validation`, `orphaned`) now lower market price (1 flag → implied 0.42; 2+ flags → implied 0.39); closes Senate→Exchange governance feedback loop\n- New REST endpoints: `POST /api/markets/artifact-support-reprice`, `GET /api/economy/artifact-support-signals`, `POST /api/markets/gate-flags-reprice`, `GET /api/economy/gate-flags-signals`\n\n### Senate Contributor Network\n- **Contribution & nomination endpoints** added to `/network/` router: `POST /network/contributions`, `GET /network/contributions`, `POST /network/nominations`, `GET /network/nominations`\n- 13 new endpoint tests; integrates with existing agent_contributions and nomination credit pipelines\n- **Activity & leaderboard endpoints**: `GET /network/agents/{id}/activity` (unified contribution feed per agent), `GET /network/leaderboard` (top contributors by trust score/contributions); 20 new tests\n- **Contribution approval workflow**: `PATCH /network/contributions/{id}` for status transitions (approved/rejected/needs_revision); auto-promotion logic fires when thresholds met\n- **Ecosystem overview route**: `GET /network/ecosystem` — aggregate health dashboard for governance dashboards; covers active contributors, squad health, recent activity\n\n### Senate: Test Coverage & Code Quality\n- **Senate test suite expanded to 130 tests** across contributor network, approval workflow, auto-promotion, and ecosystem routes\n- 15 new tests for contribution approval, auto-promotion from probation, and ecosystem health checks\n- **`datetime.utcnow()` deprecation fixed** across senate codebase (replaced with `datetime.now(timezone.utc)`)\n- **Contributor network DB path fallback hardened**: `_get_db_path()` simplified to check local path first; `include_registry` parameter now reliably returns all 20 registered agents\n\n### Exchange: Squad Harvest Driver\n- **Driver #22** (squad harvest cycle) operational with cron setup at every-1h schedule\n- Harvests mature research squads; cron configured via `setup_squad_harvest_cron.sh`\n\n### Economics: Nomination Processor Bug Fix\n- **`agent_nomination_processor.py`** expanded `downstream_status` filter from `'queued'` to include `'done'`, `'completed'`, `'agenda_debate'`, `'governance_pending'`\n- Previously, settled nominations with any status other than `'queued'` were silently skipped and never credited; fix ensures all settled nominations are processed\n\n### Exchange: Price History Snapshot\n- **355 hypothesis prices** snapshotted into price history tables; establishes baseline time-series for trend and velocity signals\n- Price history now meaningful for all major hypotheses; volatility and momentum indicators operational\n\n### Documentation System\n- Full `/docs` section with 22 core docs + 51 AI-tool pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 22 core docs verified working across `/docs/*` routes\n- Reference docs expanded: glossary (199→759 words), changelog (215→671 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (749 hyps, 396 analyses, 711711 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,664 total artifacts across 17 types; 242 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **118 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (17,406 pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–13)\n\n### System Scale (This Cycle)\n- **984 hypotheses** now active\n- **396 analyses** completed\n- **711,721 knowledge edges** in the graph\n- **3,374/3,383 open knowledge gaps** identified\n\n\n### Senate: Integrity Sweeper (April 21)\n- **New Senate integrity sweeper** detects and repairs broken links, orphan artifacts, and duplicate entities\n- `POST /api/senate/integrity_sweeper/run` triggers the sweep; `GET /api/senate/integrity_sweeper/status` monitors progress\n- Automated repair: broken link URL normalization, orphan artifact reparenting, duplicate entity consolidation\n- Audit trail via `GET /api/senate/integrity_sweeper/audit_log`\n\n### Exchange Pricing Evolution (WS33)\n- **17th pricing dimension** added: Reproducibility Score — `reproducibility_score` (range 0.10–0.85, avg 0.561, 93% coverage) measures how reliably the experimental evidence behind a hypothesis can be replicated across labs and conditions; formula: `repro_implied_price = 0.446 + normalized × 0.12` → range [0.446, 0.566]; partial Pearson r=+0.226 after controlling for composite_score (genuinely independent signal — Ioannidis replication crisis makes reproducibility first-class scientific quality criterion)\n- 32/353 hypotheses repriced on first run; first-run results: \"Closed-loop transcranial focused ultrasound with 40Hz g\" (0.668→0.663), \"Astrocytic-Mediated Tau Clearance Dysfunction via TREM2\" (0.457→0.461), \"Alpha-gamma cross-frequency coupling enhancement\" (0.627→0.624)\n- New REST endpoints: `POST /api/markets/reproducibility-reprice`, `GET /api/economy/reproducibility-signals`\n- Consumer loop: `cycle % 3024 == 0` (~50.4h cadence — between WS32 at 48h and WS29 at 56h); idempotent (72h window, 1% noise floor)\n- `\"reproducibility_signal\": \"Reproducibility\"` added to `_SIGNAL_LABELS_MAP`\n- **ACTIVE_EVENTS staleness fix**: `participant_signal` and `reproducibility_signal` added to `ACTIVE_EVENTS` tuple in `apply_staleness_decay_batch()` so market-participant evaluations properly reset the staleness clock (93 events every ~6h)\n\n### Exchange Pricing Evolution (WS31)\n- **15th pricing dimension** added: Competitive Landscape Novelty — `competitive_landscape_score` (0.1–1.0) signals first-in-class novelty vs crowded/me-too; high score pushes price toward 0.60, low score discounts toward 0.42; Pearson r=+0.123 (uncaptured before WS31, improved to r=0.149 after first run)\n- 272/332 hypotheses repriced on first run; first-in-class hypotheses (score=1.0) that were severely underpriced corrected upward\n- New REST endpoints: `POST /api/markets/competitive-landscape-reprice`, `GET /api/economy/competitive-landscape-signals`\n- Consumer loop: every 2,520 cycles (~42h) in `api.py`; idempotent (72h window, 1% noise floor)\n- Signal label \"Competitive Landscape\" added to `_SIGNAL_LABELS_MAP`; `competitive_landscape_signal` added to `ACTIVE_EVENTS`\n\n### Exchange: Agent Market Auto-Creation\n- **Agent markets auto-created on registration** — every new agent now gets a `markets` row automatically when `register_agent()` is called; no manual backfill step needed\n- `backfill_agent_markets()` utility added to `agent_registry.py` for idempotent catch-up; backfilled 11 missing markets for existing active agents\n- Market row uses `market_type='agent'`, `entity_id=<agent_id>`, initial price set from `reputation_score` (or default 50.0); closes the loop so all contributors appear in Exchange\n\n### Senate: Package Refactor (scidex/ Layer Subpackages)\n- **Codebase reorganized into proper Python package** matching the five-layer architecture: `scidex/core/`, `scidex/agora/`, `scidex/exchange/`, `scidex/forge/`, `scidex/atlas/`, `scidex/senate/`\n- 69 library modules moved under `scidex/` subpackages; each layer now has a dedicated namespace\n- **Backward compatibility preserved**: 69 thin re-export shims at root level (`import database` still works alongside `from scidex.core import database`)\n- Agents using `from <module> import ...` syntax see no change; agents may gradually migrate to `from scidex.<layer> import <module>` for clarity\n- Root-level `scripts/` directory organized: core scripts retained in `scripts/`, archived/dead scripts moved to `scripts/archive/oneoff_scripts/`\n\n### Exchange Pricing Evolution (WS30)\n- **14th pricing dimension** added: KG Artifact Support Density — counts inbound `artifact_links` (supports/derives_from/experiment_target) per hypothesis; normalized via `1-exp(-count/8)` → price range [0.44, 0.58]; Pearson r=+0.044 vs prior price (genuinely uncaptured signal)\n- **Senate gate-flags market penalty**: active quality flags (`low_validation`, `orphaned`) now lower market price (1 flag → implied 0.42; 2+ flags → implied 0.39); closes Senate→Exchange governance feedback loop\n- New REST endpoints: `POST /api/markets/artifact-support-reprice`, `GET /api/economy/artifact-support-signals`, `POST /api/markets/gate-flags-reprice`, `GET /api/economy/gate-flags-signals`\n\n### Senate Contributor Network\n- **Contribution & nomination endpoints** added to `/network/` router: `POST /network/contributions`, `GET /network/contributions`, `POST /network/nominations`, `GET /network/nominations`\n- 13 new endpoint tests; integrates with existing agent_contributions and nomination credit pipelines\n- **Activity & leaderboard endpoints**: `GET /network/agents/{id}/activity` (unified contribution feed per agent), `GET /network/leaderboard` (top contributors by trust score/contributions); 20 new tests\n- **Contribution approval workflow**: `PATCH /network/contributions/{id}` for status transitions (approved/rejected/needs_revision); auto-promotion logic fires when thresholds met\n- **Ecosystem overview route**: `GET /network/ecosystem` — aggregate health dashboard for governance dashboards; covers active contributors, squad health, recent activity\n\n### Senate: Test Coverage & Code Quality\n- **Senate test suite expanded to 130 tests** across contributor network, approval workflow, auto-promotion, and ecosystem routes\n- 15 new tests for contribution approval, auto-promotion from probation, and ecosystem health checks\n- **`datetime.utcnow()` deprecation fixed** across senate codebase (replaced with `datetime.now(timezone.utc)`)\n- **Contributor network DB path fallback hardened**: `_get_db_path()` simplified to check local path first; `include_registry` parameter now reliably returns all 20 registered agents\n\n### Exchange: Squad Harvest Driver\n- **Driver #22** (squad harvest cycle) operational with cron setup at every-1h schedule\n- Harvests mature research squads; cron configured via `setup_squad_harvest_cron.sh`\n\n### Economics: Nomination Processor Bug Fix\n- **`agent_nomination_processor.py`** expanded `downstream_status` filter from `'queued'` to include `'done'`, `'completed'`, `'agenda_debate'`, `'governance_pending'`\n- Previously, settled nominations with any status other than `'queued'` were silently skipped and never credited; fix ensures all settled nominations are processed\n\n### Exchange: Price History Snapshot\n- **355 hypothesis prices** snapshotted into price history tables; establishes baseline time-series for trend and velocity signals\n- Price history now meaningful for all major hypotheses; volatility and momentum indicators operational\n\n### Documentation System\n- Full `/docs` section with 22 core docs + 51 AI-tool pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 22 core docs verified working across `/docs/*` routes\n- Reference docs expanded: glossary (199→759 words), changelog (215→671 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (749 hyps, 396 analyses, 711,721 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,664 total artifacts across 17 types; 242 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **118 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (17,406 pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–13)\n\n### System Scale (This Cycle)\n- **984 hypotheses** now active\n- **396 analyses** completed\n- **711,721 knowledge edges** in the graph\n- **3,374/3,374 open knowledge gaps** identified\n\n### Exchange Pricing Evolution (WS33)\n- **17th pricing dimension** added: Reproducibility Score — `reproducibility_score` (range 0.10–0.85, avg 0.561, 93% coverage) measures how reliably the experimental evidence behind a hypothesis can be replicated across labs and conditions; formula: `repro_implied_price = 0.446 + normalized × 0.12` → range [0.446, 0.566]; partial Pearson r=+0.226 after controlling for composite_score (genuinely independent signal — Ioannidis replication crisis makes reproducibility first-class scientific quality criterion)\n- 32/353 hypotheses repriced on first run; first-run results: \"Closed-loop transcranial focused ultrasound with 40Hz g\" (0.668→0.663), \"Astrocytic-Mediated Tau Clearance Dysfunction via TREM2\" (0.457→0.461), \"Alpha-gamma cross-frequency coupling enhancement\" (0.627→0.624)\n- New REST endpoints: `POST /api/markets/reproducibility-reprice`, `GET /api/economy/reproducibility-signals`\n- Consumer loop: `cycle % 3024 == 0` (~50.4h cadence — between WS32 at 48h and WS29 at 56h); idempotent (72h window, 1% noise floor)\n- `\"reproducibility_signal\": \"Reproducibility\"` added to `_SIGNAL_LABELS_MAP`\n- **ACTIVE_EVENTS staleness fix**: `participant_signal` and `reproducibility_signal` added to `ACTIVE_EVENTS` tuple in `apply_staleness_decay_batch()` so market-participant evaluations properly reset the staleness clock (93 events every ~6h)\n\n### Exchange Pricing Evolution (WS31)\n- **15th pricing dimension** added: Competitive Landscape Novelty — `competitive_landscape_score` (0.1–1.0) signals first-in-class novelty vs crowded/me-too; high score pushes price toward 0.60, low score discounts toward 0.42; Pearson r=+0.123 (uncaptured before WS31, improved to r=0.149 after first run)\n- 272/332 hypotheses repriced on first run; first-in-class hypotheses (score=1.0) that were severely underpriced corrected upward\n- New REST endpoints: `POST /api/markets/competitive-landscape-reprice`, `GET /api/economy/competitive-landscape-signals`\n- Consumer loop: every 2,520 cycles (~42h) in `api.py`; idempotent (72h window, 1% noise floor)\n- Signal label \"Competitive Landscape\" added to `_SIGNAL_LABELS_MAP`; `competitive_landscape_signal` added to `ACTIVE_EVENTS`\n\n### Exchange: Agent Market Auto-Creation\n- **Agent markets auto-created on registration** — every new agent now gets a `markets` row automatically when `register_agent()` is called; no manual backfill step needed\n- `backfill_agent_markets()` utility added to `agent_registry.py` for idempotent catch-up; backfilled 11 missing markets for existing active agents\n- Market row uses `market_type='agent'`, `entity_id=<agent_id>`, initial price set from `reputation_score` (or default 50.0); closes the loop so all contributors appear in Exchange\n\n### Senate: Package Refactor (scidex/ Layer Subpackages)\n- **Codebase reorganized into proper Python package** matching the five-layer architecture: `scidex/core/`, `scidex/agora/`, `scidex/exchange/`, `scidex/forge/`, `scidex/atlas/`, `scidex/senate/`\n- 69 library modules moved under `scidex/` subpackages; each layer now has a dedicated namespace\n- **Backward compatibility preserved**: 69 thin re-export shims at root level (`import database` still works alongside `from scidex.core import database`)\n- Agents using `from <module> import ...` syntax see no change; agents may gradually migrate to `from scidex.<layer> import <module>` for clarity\n- Root-level `scripts/` directory organized: core scripts retained in `scripts/`, archived/dead scripts moved to `scripts/archive/oneoff_scripts/`\n\n### Exchange Pricing Evolution (WS30)\n- **14th pricing dimension** added: KG Artifact Support Density — counts inbound `artifact_links` (supports/derives_from/experiment_target) per hypothesis; normalized via `1-exp(-count/8)` → price range [0.44, 0.58]; Pearson r=+0.044 vs prior price (genuinely uncaptured signal)\n- **Senate gate-flags market penalty**: active quality flags (`low_validation`, `orphaned`) now lower market price (1 flag → implied 0.42; 2+ flags → implied 0.39); closes Senate→Exchange governance feedback loop\n- New REST endpoints: `POST /api/markets/artifact-support-reprice`, `GET /api/economy/artifact-support-signals`, `POST /api/markets/gate-flags-reprice`, `GET /api/economy/gate-flags-signals`\n\n### Senate Contributor Network\n- **Contribution & nomination endpoints** added to `/network/` router: `POST /network/contributions`, `GET /network/contributions`, `POST /network/nominations`, `GET /network/nominations`\n- 13 new endpoint tests; integrates with existing agent_contributions and nomination credit pipelines\n- **Activity & leaderboard endpoints**: `GET /network/agents/{id}/activity` (unified contribution feed per agent), `GET /network/leaderboard` (top contributors by trust score/contributions); 20 new tests\n- **Contribution approval workflow**: `PATCH /network/contributions/{id}` for status transitions (approved/rejected/needs_revision); auto-promotion logic fires when thresholds met\n- **Ecosystem overview route**: `GET /network/ecosystem` — aggregate health dashboard for governance dashboards; covers active contributors, squad health, recent activity\n\n### Senate: Test Coverage & Code Quality\n- **Senate test suite expanded to 130 tests** across contributor network, approval workflow, auto-promotion, and ecosystem routes\n- 15 new tests for contribution approval, auto-promotion from probation, and ecosystem health checks\n- **`datetime.utcnow()` deprecation fixed** across senate codebase (replaced with `datetime.now(timezone.utc)`)\n- **Contributor network DB path fallback hardened**: `_get_db_path()` simplified to check local path first; `include_registry` parameter now reliably returns all 20 registered agents\n\n### Exchange: Squad Harvest Driver\n- **Driver #22** (squad harvest cycle) operational with cron setup at every-1h schedule\n- Harvests mature research squads; cron configured via `setup_squad_harvest_cron.sh`\n\n### Economics: Nomination Processor Bug Fix\n- **`agent_nomination_processor.py`** expanded `downstream_status` filter from `'queued'` to include `'done'`, `'completed'`, `'agenda_debate'`, `'governance_pending'`\n- Previously, settled nominations with any status other than `'queued'` were silently skipped and never credited; fix ensures all settled nominations are processed\n\n### Exchange: Price History Snapshot\n- **355 hypothesis prices** snapshotted into price history tables; establishes baseline time-series for trend and velocity signals\n- Price history now meaningful for all major hypotheses; volatility and momentum indicators operational\n\n### Documentation System\n- Full `/docs` section with 22 core docs + 51 AI-tool pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 22 core docs verified working across `/docs/*` routes\n- Reference docs expanded: glossary (199→759 words), changelog (215→671 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (749 hyps, 396 analyses, 711,721 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,664 total artifacts across 17 types; 242 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **118 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (17,406 pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–13)\n\n### System Scale (This Cycle)\n- **984 hypotheses** now active\n- **396 analyses** completed\n- **711,721 knowledge edges** in the graph\n- **3,374/3,383 open knowledge gaps** identified\n\n### Exchange Pricing Evolution (WS33)\n- **17th pricing dimension** added: Reproducibility Score — `reproducibility_score` (range 0.10–0.85, avg 0.561, 93% coverage) measures how reliably the experimental evidence behind a hypothesis can be replicated across labs and conditions; formula: `repro_implied_price = 0.446 + normalized × 0.12` → range [0.446, 0.566]; partial Pearson r=+0.226 after controlling for composite_score (genuinely independent signal — Ioannidis replication crisis makes reproducibility first-class scientific quality criterion)\n- 32/353 hypotheses repriced on first run; first-run results: \"Closed-loop transcranial focused ultrasound with 40Hz g\" (0.668→0.663), \"Astrocytic-Mediated Tau Clearance Dysfunction via TREM2\" (0.457→0.461), \"Alpha-gamma cross-frequency coupling enhancement\" (0.627→0.624)\n- New REST endpoints: `POST /api/markets/reproducibility-reprice`, `GET /api/economy/reproducibility-signals`\n- Consumer loop: `cycle % 3024 == 0` (~50.4h cadence — between WS32 at 48h and WS29 at 56h); idempotent (72h window, 1% noise floor)\n- `\"reproducibility_signal\": \"Reproducibility\"` added to `_SIGNAL_LABELS_MAP`\n- **ACTIVE_EVENTS staleness fix**: `participant_signal` and `reproducibility_signal` added to `ACTIVE_EVENTS` tuple in `apply_staleness_decay_batch()` so market-participant evaluations properly reset the staleness clock (93 events every ~6h)\n\n### Exchange Pricing Evolution (WS31)\n- **15th pricing dimension** added: Competitive Landscape Novelty — `competitive_landscape_score` (0.1–1.0) signals first-in-class novelty vs crowded/me-too; high score pushes price toward 0.60, low score discounts toward 0.42; Pearson r=+0.123 (uncaptured before WS31, improved to r=0.149 after first run)\n- 272/332 hypotheses repriced on first run; first-in-class hypotheses (score=1.0) that were severely underpriced corrected upward\n- New REST endpoints: `POST /api/markets/competitive-landscape-reprice`, `GET /api/economy/competitive-landscape-signals`\n- Consumer loop: every 2,520 cycles (~42h) in `api.py`; idempotent (72h window, 1% noise floor)\n- Signal label \"Competitive Landscape\" added to `_SIGNAL_LABELS_MAP`; `competitive_landscape_signal` added to `ACTIVE_EVENTS`\n\n### Exchange: Agent Market Auto-Creation\n- **Agent markets auto-created on registration** — every new agent now gets a `markets` row automatically when `register_agent()` is called; no manual backfill step needed\n- `backfill_agent_markets()` utility added to `agent_registry.py` for idempotent catch-up; backfilled 11 missing markets for existing active agents\n- Market row uses `market_type='agent'`, `entity_id=<agent_id>`, initial price set from `reputation_score` (or default 50.0); closes the loop so all contributors appear in Exchange\n\n### Senate: Package Refactor (scidex/ Layer Subpackages)\n- **Codebase reorganized into proper Python package** matching the five-layer architecture: `scidex/core/`, `scidex/agora/`, `scidex/exchange/`, `scidex/forge/`, `scidex/atlas/`, `scidex/senate/`\n- 69 library modules moved under `scidex/` subpackages; each layer now has a dedicated namespace\n- **Backward compatibility preserved**: 69 thin re-export shims at root level (`import database` still works alongside `from scidex.core import database`)\n- Agents using `from <module> import ...` syntax see no change; agents may gradually migrate to `from scidex.<layer> import <module>` for clarity\n- Root-level `scripts/` directory organized: core scripts retained in `scripts/`, archived/dead scripts moved to `scripts/archive/oneoff_scripts/`\n\n### Exchange Pricing Evolution (WS30)\n- **14th pricing dimension** added: KG Artifact Support Density — counts inbound `artifact_links` (supports/derives_from/experiment_target) per hypothesis; normalized via `1-exp(-count/8)` → price range [0.44, 0.58]; Pearson r=+0.044 vs prior price (genuinely uncaptured signal)\n- **Senate gate-flags market penalty**: active quality flags (`low_validation`, `orphaned`) now lower market price (1 flag → implied 0.42; 2+ flags → implied 0.39); closes Senate→Exchange governance feedback loop\n- New REST endpoints: `POST /api/markets/artifact-support-reprice`, `GET /api/economy/artifact-support-signals`, `POST /api/markets/gate-flags-reprice`, `GET /api/economy/gate-flags-signals`\n\n### Senate Contributor Network\n- **Contribution & nomination endpoints** added to `/network/` router: `POST /network/contributions`, `GET /network/contributions`, `POST /network/nominations`, `GET /network/nominations`\n- 13 new endpoint tests; integrates with existing agent_contributions and nomination credit pipelines\n- **Activity & leaderboard endpoints**: `GET /network/agents/{id}/activity` (unified contribution feed per agent), `GET /network/leaderboard` (top contributors by trust score/contributions); 20 new tests\n- **Contribution approval workflow**: `PATCH /network/contributions/{id}` for status transitions (approved/rejected/needs_revision); auto-promotion logic fires when thresholds met\n- **Ecosystem overview route**: `GET /network/ecosystem` — aggregate health dashboard for governance dashboards; covers active contributors, squad health, recent activity\n\n### Senate: Test Coverage & Code Quality\n- **Senate test suite expanded to 130 tests** across contributor network, approval workflow, auto-promotion, and ecosystem routes\n- 15 new tests for contribution approval, auto-promotion from probation, and ecosystem health checks\n- **`datetime.utcnow()` deprecation fixed** across senate codebase (replaced with `datetime.now(timezone.utc)`)\n- **Contributor network DB path fallback hardened**: `_get_db_path()` simplified to check local path first; `include_registry` parameter now reliably returns all 20 registered agents\n\n### Exchange: Squad Harvest Driver\n- **Driver #22** (squad harvest cycle) operational with cron setup at every-1h schedule\n- Harvests mature research squads; cron configured via `setup_squad_harvest_cron.sh`\n\n### Economics: Nomination Processor Bug Fix\n- **`agent_nomination_processor.py`** expanded `downstream_status` filter from `'queued'` to include `'done'`, `'completed'`, `'agenda_debate'`, `'governance_pending'`\n- Previously, settled nominations with any status other than `'queued'` were silently skipped and never credited; fix ensures all settled nominations are processed\n\n### Exchange: Price History Snapshot\n- **355 hypothesis prices** snapshotted into price history tables; establishes baseline time-series for trend and velocity signals\n- Price history now meaningful for all major hypotheses; volatility and momentum indicators operational\n\n### Documentation System\n- Full `/docs` section with 22 core docs + 51 AI-tool pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 22 core docs verified working across `/docs/*` routes\n- Reference docs expanded: glossary (199→759 words), changelog (215→671 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (749 hyps, 396 analyses, 711,721 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,664 total artifacts across 17 types; 242 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **118 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (17,406 pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–13)\n\n### System Scale (This Cycle)\n- **749 hypotheses** now active\n- **396 analyses** completed\n- **711,721 knowledge edges** in the graph\n- **3,372/3,383 open knowledge gaps** identified\n\n### Exchange Pricing Evolution (WS33)\n- **17th pricing dimension** added: Reproducibility Score — `reproducibility_score` (range 0.10–0.85, avg 0.561, 93% coverage) measures how reliably the experimental evidence behind a hypothesis can be replicated across labs and conditions; formula: `repro_implied_price = 0.446 + normalized × 0.12` → range [0.446, 0.566]; partial Pearson r=+0.226 after controlling for composite_score (genuinely independent signal — Ioannidis replication crisis makes reproducibility first-class scientific quality criterion)\n- 32/353 hypotheses repriced on first run; first-run results: \"Closed-loop transcranial focused ultrasound with 40Hz g\" (0.668→0.663), \"Astrocytic-Mediated Tau Clearance Dysfunction via TREM2\" (0.457→0.461), \"Alpha-gamma cross-frequency coupling enhancement\" (0.627→0.624)\n- New REST endpoints: `POST /api/markets/reproducibility-reprice`, `GET /api/economy/reproducibility-signals`\n- Consumer loop: `cycle % 3024 == 0` (~50.4h cadence — between WS32 at 48h and WS29 at 56h); idempotent (72h window, 1% noise floor)\n- `\"reproducibility_signal\": \"Reproducibility\"` added to `_SIGNAL_LABELS_MAP`\n- **ACTIVE_EVENTS staleness fix**: `participant_signal` and `reproducibility_signal` added to `ACTIVE_EVENTS` tuple in `apply_staleness_decay_batch()` so market-participant evaluations properly reset the staleness clock (93 events every ~6h)\n\n### Exchange Pricing Evolution (WS31)\n- **15th pricing dimension** added: Competitive Landscape Novelty — `competitive_landscape_score` (0.1–1.0) signals first-in-class novelty vs crowded/me-too; high score pushes price toward 0.60, low score discounts toward 0.42; Pearson r=+0.123 (uncaptured before WS31, improved to r=0.149 after first run)\n- 272/332 hypotheses repriced on first run; first-in-class hypotheses (score=1.0) that were severely underpriced corrected upward\n- New REST endpoints: `POST /api/markets/competitive-landscape-reprice`, `GET /api/economy/competitive-landscape-signals`\n- Consumer loop: every 2,520 cycles (~42h) in `api.py`; idempotent (72h window, 1% noise floor)\n- Signal label \"Competitive Landscape\" added to `_SIGNAL_LABELS_MAP`; `competitive_landscape_signal` added to `ACTIVE_EVENTS`\n\n### Exchange: Agent Market Auto-Creation\n- **Agent markets auto-created on registration** — every new agent now gets a `markets` row automatically when `register_agent()` is called; no manual backfill step needed\n- `backfill_agent_markets()` utility added to `agent_registry.py` for idempotent catch-up; backfilled 11 missing markets for existing active agents\n- Market row uses `market_type='agent'`, `entity_id=<agent_id>`, initial price set from `reputation_score` (or default 50.0); closes the loop so all contributors appear in Exchange\n\n### Senate: Package Refactor (scidex/ Layer Subpackages)\n- **Codebase reorganized into proper Python package** matching the five-layer architecture: `scidex/core/`, `scidex/agora/`, `scidex/exchange/`, `scidex/forge/`, `scidex/atlas/`, `scidex/senate/`\n- 69 library modules moved under `scidex/` subpackages; each layer now has a dedicated namespace\n- **Backward compatibility preserved**: 69 thin re-export shims at root level (`import database` still works alongside `from scidex.core import database`)\n- Agents using `from <module> import ...` syntax see no change; agents may gradually migrate to `from scidex.<layer> import <module>` for clarity\n- Root-level `scripts/` directory organized: core scripts retained in `scripts/`, archived/dead scripts moved to `scripts/archive/oneoff_scripts/`\n\n### Exchange Pricing Evolution (WS30)\n- **14th pricing dimension** added: KG Artifact Support Density — counts inbound `artifact_links` (supports/derives_from/experiment_target) per hypothesis; normalized via `1-exp(-count/8)` → price range [0.44, 0.58]; Pearson r=+0.044 vs prior price (genuinely uncaptured signal)\n- **Senate gate-flags market penalty**: active quality flags (`low_validation`, `orphaned`) now lower market price (1 flag → implied 0.42; 2+ flags → implied 0.39); closes Senate→Exchange governance feedback loop\n- New REST endpoints: `POST /api/markets/artifact-support-reprice`, `GET /api/economy/artifact-support-signals`, `POST /api/markets/gate-flags-reprice`, `GET /api/economy/gate-flags-signals`\n\n### Senate Contributor Network\n- **Contribution & nomination endpoints** added to `/network/` router: `POST /network/contributions`, `GET /network/contributions`, `POST /network/nominations`, `GET /network/nominations`\n- 13 new endpoint tests; integrates with existing agent_contributions and nomination credit pipelines\n- **Activity & leaderboard endpoints**: `GET /network/agents/{id}/activity` (unified contribution feed per agent), `GET /network/leaderboard` (top contributors by trust score/contributions); 20 new tests\n- **Contribution approval workflow**: `PATCH /network/contributions/{id}` for status transitions (approved/rejected/needs_revision); auto-promotion logic fires when thresholds met\n- **Ecosystem overview route**: `GET /network/ecosystem` — aggregate health dashboard for governance dashboards; covers active contributors, squad health, recent activity\n\n### Senate: Test Coverage & Code Quality\n- **Senate test suite expanded to 130 tests** across contributor network, approval workflow, auto-promotion, and ecosystem routes\n- 15 new tests for contribution approval, auto-promotion from probation, and ecosystem health checks\n- **`datetime.utcnow()` deprecation fixed** across senate codebase (replaced with `datetime.now(timezone.utc)`)\n- **Contributor network DB path fallback hardened**: `_get_db_path()` simplified to check local path first; `include_registry` parameter now reliably returns all 20 registered agents\n\n### Exchange: Squad Harvest Driver\n- **Driver #22** (squad harvest cycle) operational with cron setup at every-1h schedule\n- Harvests mature research squads; cron configured via `setup_squad_harvest_cron.sh`\n\n### Economics: Nomination Processor Bug Fix\n- **`agent_nomination_processor.py`** expanded `downstream_status` filter from `'queued'` to include `'done'`, `'completed'`, `'agenda_debate'`, `'governance_pending'`\n- Previously, settled nominations with any status other than `'queued'` were silently skipped and never credited; fix ensures all settled nominations are processed\n\n### Exchange: Price History Snapshot\n- **355 hypothesis prices** snapshotted into price history tables; establishes baseline time-series for trend and velocity signals\n- Price history now meaningful for all major hypotheses; volatility and momentum indicators operational\n\n### Documentation System\n- Full `/docs` section with 22 core docs + 51 AI-tool pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 22 core docs verified working across `/docs/*` routes\n- Reference docs expanded: glossary (199→759 words), changelog (215→671 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (749 hyps, 396 analyses, 711,721 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,664 total artifacts across 17 types; 242 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **118 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (17,406 pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–13)\n\n### System Scale (This Cycle)\n- **984 hypotheses** now active\n- **396 analyses** completed\n- **711,721 knowledge edges** in the graph\n- **3,372/3,383 open knowledge gaps** identified\n\n### Exchange Pricing Evolution (WS33)\n- **17th pricing dimension** added: Reproducibility Score — `reproducibility_score` (range 0.10–0.85, avg 0.561, 93% coverage) measures how reliably the experimental evidence behind a hypothesis can be replicated across labs and conditions; formula: `repro_implied_price = 0.446 + normalized × 0.12` → range [0.446, 0.566]; partial Pearson r=+0.226 after controlling for composite_score (genuinely independent signal — Ioannidis replication crisis makes reproducibility first-class scientific quality criterion)\n- 32/353 hypotheses repriced on first run; first-run results: \"Closed-loop transcranial focused ultrasound with 40Hz g\" (0.668→0.663), \"Astrocytic-Mediated Tau Clearance Dysfunction via TREM2\" (0.457→0.461), \"Alpha-gamma cross-frequency coupling enhancement\" (0.627→0.624)\n- New REST endpoints: `POST /api/markets/reproducibility-reprice`, `GET /api/economy/reproducibility-signals`\n- Consumer loop: `cycle % 3024 == 0` (~50.4h cadence — between WS32 at 48h and WS29 at 56h); idempotent (72h window, 1% noise floor)\n- `\"reproducibility_signal\": \"Reproducibility\"` added to `_SIGNAL_LABELS_MAP`\n- **ACTIVE_EVENTS staleness fix**: `participant_signal` and `reproducibility_signal` added to `ACTIVE_EVENTS` tuple in `apply_staleness_decay_batch()` so market-participant evaluations properly reset the staleness clock (93 events every ~6h)\n\n### Exchange Pricing Evolution (WS31)\n- **15th pricing dimension** added: Competitive Landscape Novelty — `competitive_landscape_score` (0.1–1.0) signals first-in-class novelty vs crowded/me-too; high score pushes price toward 0.60, low score discounts toward 0.42; Pearson r=+0.123 (uncaptured before WS31, improved to r=0.149 after first run)\n- 272/332 hypotheses repriced on first run; first-in-class hypotheses (score=1.0) that were severely underpriced corrected upward\n- New REST endpoints: `POST /api/markets/competitive-landscape-reprice`, `GET /api/economy/competitive-landscape-signals`\n- Consumer loop: every 2,520 cycles (~42h) in `api.py`; idempotent (72h window, 1% noise floor)\n- Signal label \"Competitive Landscape\" added to `_SIGNAL_LABELS_MAP`; `competitive_landscape_signal` added to `ACTIVE_EVENTS`\n\n### Exchange: Agent Market Auto-Creation\n- **Agent markets auto-created on registration** — every new agent now gets a `markets` row automatically when `register_agent()` is called; no manual backfill step needed\n- `backfill_agent_markets()` utility added to `agent_registry.py` for idempotent catch-up; backfilled 11 missing markets for existing active agents\n- Market row uses `market_type='agent'`, `entity_id=<agent_id>`, initial price set from `reputation_score` (or default 50.0); closes the loop so all contributors appear in Exchange\n\n### Senate: Package Refactor (scidex/ Layer Subpackages)\n- **Codebase reorganized into proper Python package** matching the five-layer architecture: `scidex/core/`, `scidex/agora/`, `scidex/exchange/`, `scidex/forge/`, `scidex/atlas/`, `scidex/senate/`\n- 69 library modules moved under `scidex/` subpackages; each layer now has a dedicated namespace\n- **Backward compatibility preserved**: 69 thin re-export shims at root level (`import database` still works alongside `from scidex.core import database`)\n- Agents using `from <module> import ...` syntax see no change; agents may gradually migrate to `from scidex.<layer> import <module>` for clarity\n- Root-level `scripts/` directory organized: core scripts retained in `scripts/`, archived/dead scripts moved to `scripts/archive/oneoff_scripts/`\n\n### Exchange Pricing Evolution (WS30)\n- **14th pricing dimension** added: KG Artifact Support Density — counts inbound `artifact_links` (supports/derives_from/experiment_target) per hypothesis; normalized via `1-exp(-count/8)` → price range [0.44, 0.58]; Pearson r=+0.044 vs prior price (genuinely uncaptured signal)\n- **Senate gate-flags market penalty**: active quality flags (`low_validation`, `orphaned`) now lower market price (1 flag → implied 0.42; 2+ flags → implied 0.39); closes Senate→Exchange governance feedback loop\n- New REST endpoints: `POST /api/markets/artifact-support-reprice`, `GET /api/economy/artifact-support-signals`, `POST /api/markets/gate-flags-reprice`, `GET /api/economy/gate-flags-signals`\n\n### Senate Contributor Network\n- **Contribution & nomination endpoints** added to `/network/` router: `POST /network/contributions`, `GET /network/contributions`, `POST /network/nominations`, `GET /network/nominations`\n- 13 new endpoint tests; integrates with existing agent_contributions and nomination credit pipelines\n- **Activity & leaderboard endpoints**: `GET /network/agents/{id}/activity` (unified contribution feed per agent), `GET /network/leaderboard` (top contributors by trust score/contributions); 20 new tests\n- **Contribution approval workflow**: `PATCH /network/contributions/{id}` for status transitions (approved/rejected/needs_revision); auto-promotion logic fires when thresholds met\n- **Ecosystem overview route**: `GET /network/ecosystem` — aggregate health dashboard for governance dashboards; covers active contributors, squad health, recent activity\n\n### Senate: Test Coverage & Code Quality\n- **Senate test suite expanded to 130 tests** across contributor network, approval workflow, auto-promotion, and ecosystem routes\n- 15 new tests for contribution approval, auto-promotion from probation, and ecosystem health checks\n- **`datetime.utcnow()` deprecation fixed** across senate codebase (replaced with `datetime.now(timezone.utc)`)\n- **Contributor network DB path fallback hardened**: `_get_db_path()` simplified to check local path first; `include_registry` parameter now reliably returns all 20 registered agents\n\n### Exchange: Squad Harvest Driver\n- **Driver #22** (squad harvest cycle) operational with cron setup at every-1h schedule\n- Harvests mature research squads; cron configured via `setup_squad_harvest_cron.sh`\n\n### Economics: Nomination Processor Bug Fix\n- **`agent_nomination_processor.py`** expanded `downstream_status` filter from `'queued'` to include `'done'`, `'completed'`, `'agenda_debate'`, `'governance_pending'`\n- Previously, settled nominations with any status other than `'queued'` were silently skipped and never credited; fix ensures all settled nominations are processed\n\n### Exchange: Price History Snapshot\n- **355 hypothesis prices** snapshotted into price history tables; establishes baseline time-series for trend and velocity signals\n- Price history now meaningful for all major hypotheses; volatility and momentum indicators operational\n\n### Documentation System\n- Full `/docs` section with 22 core docs + 51 AI-tool pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 22 core docs verified working across `/docs/*` routes\n- Reference docs expanded: glossary (199→759 words), changelog (215→671 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (749 hyps, 396 analyses, 711,721 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,664 total artifacts across 17 types; 242 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **118 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (17,406 pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–13)\n\n### System Scale (This Cycle)\n- **749 hypotheses** now active\n- **396 analyses** completed\n- **711,721 knowledge edges** in the graph\n- **3,372/3,383 open knowledge gaps** identified\n\n### Exchange Pricing Evolution (WS33)\n- **17th pricing dimension** added: Reproducibility Score — `reproducibility_score` (range 0.10–0.85, avg 0.561, 93% coverage) measures how reliably the experimental evidence behind a hypothesis can be replicated across labs and conditions; formula: `repro_implied_price = 0.446 + normalized × 0.12` → range [0.446, 0.566]; partial Pearson r=+0.226 after controlling for composite_score (genuinely independent signal — Ioannidis replication crisis makes reproducibility first-class scientific quality criterion)\n- 32/353 hypotheses repriced on first run; first-run results: \"Closed-loop transcranial focused ultrasound with 40Hz g\" (0.668→0.663), \"Astrocytic-Mediated Tau Clearance Dysfunction via TREM2\" (0.457→0.461), \"Alpha-gamma cross-frequency coupling enhancement\" (0.627→0.624)\n- New REST endpoints: `POST /api/markets/reproducibility-reprice`, `GET /api/economy/reproducibility-signals`\n- Consumer loop: `cycle % 3024 == 0` (~50.4h cadence — between WS32 at 48h and WS29 at 56h); idempotent (72h window, 1% noise floor)\n- `\"reproducibility_signal\": \"Reproducibility\"` added to `_SIGNAL_LABELS_MAP`\n- **ACTIVE_EVENTS staleness fix**: `participant_signal` and `reproducibility_signal` added to `ACTIVE_EVENTS` tuple in `apply_staleness_decay_batch()` so market-participant evaluations properly reset the staleness clock (93 events every ~6h)\n\n### Exchange Pricing Evolution (WS31)\n- **15th pricing dimension** added: Competitive Landscape Novelty — `competitive_landscape_score` (0.1–1.0) signals first-in-class novelty vs crowded/me-too; high score pushes price toward 0.60, low score discounts toward 0.42; Pearson r=+0.123 (uncaptured before WS31, improved to r=0.149 after first run)\n- 272/332 hypotheses repriced on first run; first-in-class hypotheses (score=1.0) that were severely underpriced corrected upward\n- New REST endpoints: `POST /api/markets/competitive-landscape-reprice`, `GET /api/economy/competitive-landscape-signals`\n- Consumer loop: every 2,520 cycles (~42h) in `api.py`; idempotent (72h window, 1% noise floor)\n- Signal label \"Competitive Landscape\" added to `_SIGNAL_LABELS_MAP`; `competitive_landscape_signal` added to `ACTIVE_EVENTS`\n\n### Exchange: Agent Market Auto-Creation\n- **Agent markets auto-created on registration** — every new agent now gets a `markets` row automatically when `register_agent()` is called; no manual backfill step needed\n- `backfill_agent_markets()` utility added to `agent_registry.py` for idempotent catch-up; backfilled 11 missing markets for existing active agents\n- Market row uses `market_type='agent'`, `entity_id=<agent_id>`, initial price set from `reputation_score` (or default 50.0); closes the loop so all contributors appear in Exchange\n\n### Senate: Package Refactor (scidex/ Layer Subpackages)\n- **Codebase reorganized into proper Python package** matching the five-layer architecture: `scidex/core/`, `scidex/agora/`, `scidex/exchange/`, `scidex/forge/`, `scidex/atlas/`, `scidex/senate/`\n- 69 library modules moved under `scidex/` subpackages; each layer now has a dedicated namespace\n- **Backward compatibility preserved**: 69 thin re-export shims at root level (`import database` still works alongside `from scidex.core import database`)\n- Agents using `from <module> import ...` syntax see no change; agents may gradually migrate to `from scidex.<layer> import <module>` for clarity\n- Root-level `scripts/` directory organized: core scripts retained in `scripts/`, archived/dead scripts moved to `scripts/archive/oneoff_scripts/`\n\n### Exchange Pricing Evolution (WS30)\n- **14th pricing dimension** added: KG Artifact Support Density — counts inbound `artifact_links` (supports/derives_from/experiment_target) per hypothesis; normalized via `1-exp(-count/8)` → price range [0.44, 0.58]; Pearson r=+0.044 vs prior price (genuinely uncaptured signal)\n- **Senate gate-flags market penalty**: active quality flags (`low_validation`, `orphaned`) now lower market price (1 flag → implied 0.42; 2+ flags → implied 0.39); closes Senate→Exchange governance feedback loop\n- New REST endpoints: `POST /api/markets/artifact-support-reprice`, `GET /api/economy/artifact-support-signals`, `POST /api/markets/gate-flags-reprice`, `GET /api/economy/gate-flags-signals`\n\n### Senate Contributor Network\n- **Contribution & nomination endpoints** added to `/network/` router: `POST /network/contributions`, `GET /network/contributions`, `POST /network/nominations`, `GET /network/nominations`\n- 13 new endpoint tests; integrates with existing agent_contributions and nomination credit pipelines\n- **Activity & leaderboard endpoints**: `GET /network/agents/{id}/activity` (unified contribution feed per agent), `GET /network/leaderboard` (top contributors by trust score/contributions); 20 new tests\n- **Contribution approval workflow**: `PATCH /network/contributions/{id}` for status transitions (approved/rejected/needs_revision); auto-promotion logic fires when thresholds met\n- **Ecosystem overview route**: `GET /network/ecosystem` — aggregate health dashboard for governance dashboards; covers active contributors, squad health, recent activity\n\n### Senate: Test Coverage & Code Quality\n- **Senate test suite expanded to 130 tests** across contributor network, approval workflow, auto-promotion, and ecosystem routes\n- 15 new tests for contribution approval, auto-promotion from probation, and ecosystem health checks\n- **`datetime.utcnow()` deprecation fixed** across senate codebase (replaced with `datetime.now(timezone.utc)`)\n- **Contributor network DB path fallback hardened**: `_get_db_path()` simplified to check local path first; `include_registry` parameter now reliably returns all 20 registered agents\n\n### Exchange: Squad Harvest Driver\n- **Driver #22** (squad harvest cycle) operational with cron setup at every-1h schedule\n- Harvests mature research squads; cron configured via `setup_squad_harvest_cron.sh`\n\n### Economics: Nomination Processor Bug Fix\n- **`agent_nomination_processor.py`** expanded `downstream_status` filter from `'queued'` to include `'done'`, `'completed'`, `'agenda_debate'`, `'governance_pending'`\n- Previously, settled nominations with any status other than `'queued'` were silently skipped and never credited; fix ensures all settled nominations are processed\n\n### Exchange: Price History Snapshot\n- **355 hypothesis prices** snapshotted into price history tables; establishes baseline time-series for trend and velocity signals\n- Price history now meaningful for all major hypotheses; volatility and momentum indicators operational\n\n### Documentation System\n- Full `/docs` section with 22 core docs + 51 AI-tool pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 22 core docs verified working across `/docs/*` routes\n- Reference docs expanded: glossary (199→759 words), changelog (215→671 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (749 hyps, 396 analyses, 711,721 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,664 total artifacts across 17 types; 242 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **118 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (17,406 pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–13)\n\n### System Scale (This Cycle)\n- **747 hypotheses** now active\n- **395 analyses** completed\n- **711,721 knowledge edges** in the graph\n- **3,091/3,383 open knowledge gaps** identified\n\n### Exchange Pricing Evolution (WS33)\n- **17th pricing dimension** added: Reproducibility Score — `reproducibility_score` (range 0.10–0.85, avg 0.561, 93% coverage) measures how reliably the experimental evidence behind a hypothesis can be replicated across labs and conditions; formula: `repro_implied_price = 0.446 + normalized × 0.12` → range [0.446, 0.566]; partial Pearson r=+0.226 after controlling for composite_score (genuinely independent signal — Ioannidis replication crisis makes reproducibility first-class scientific quality criterion)\n- 32/353 hypotheses repriced on first run; first-run results: \"Closed-loop transcranial focused ultrasound with 40Hz g\" (0.668→0.663), \"Astrocytic-Mediated Tau Clearance Dysfunction via TREM2\" (0.457→0.461), \"Alpha-gamma cross-frequency coupling enhancement\" (0.627→0.624)\n- New REST endpoints: `POST /api/markets/reproducibility-reprice`, `GET /api/economy/reproducibility-signals`\n- Consumer loop: `cycle % 3024 == 0` (~50.4h cadence — between WS32 at 48h and WS29 at 56h); idempotent (72h window, 1% noise floor)\n- `\"reproducibility_signal\": \"Reproducibility\"` added to `_SIGNAL_LABELS_MAP`\n- **ACTIVE_EVENTS staleness fix**: `participant_signal` and `reproducibility_signal` added to `ACTIVE_EVENTS` tuple in `apply_staleness_decay_batch()` so market-participant evaluations properly reset the staleness clock (93 events every ~6h)\n\n### Exchange Pricing Evolution (WS31)\n- **15th pricing dimension** added: Competitive Landscape Novelty — `competitive_landscape_score` (0.1–1.0) signals first-in-class novelty vs crowded/me-too; high score pushes price toward 0.60, low score discounts toward 0.42; Pearson r=+0.123 (uncaptured before WS31, improved to r=0.149 after first run)\n- 272/332 hypotheses repriced on first run; first-in-class hypotheses (score=1.0) that were severely underpriced corrected upward\n- New REST endpoints: `POST /api/markets/competitive-landscape-reprice`, `GET /api/economy/competitive-landscape-signals`\n- Consumer loop: every 2,520 cycles (~42h) in `api.py`; idempotent (72h window, 1% noise floor)\n- Signal label \"Competitive Landscape\" added to `_SIGNAL_LABELS_MAP`; `competitive_landscape_signal` added to `ACTIVE_EVENTS`\n\n### Exchange: Agent Market Auto-Creation\n- **Agent markets auto-created on registration** — every new agent now gets a `markets` row automatically when `register_agent()` is called; no manual backfill step needed\n- `backfill_agent_markets()` utility added to `agent_registry.py` for idempotent catch-up; backfilled 11 missing markets for existing active agents\n- Market row uses `market_type='agent'`, `entity_id=<agent_id>`, initial price set from `reputation_score` (or default 50.0); closes the loop so all contributors appear in Exchange\n\n### Senate: Package Refactor (scidex/ Layer Subpackages)\n- **Codebase reorganized into proper Python package** matching the five-layer architecture: `scidex/core/`, `scidex/agora/`, `scidex/exchange/`, `scidex/forge/`, `scidex/atlas/`, `scidex/senate/`\n- 69 library modules moved under `scidex/` subpackages; each layer now has a dedicated namespace\n- **Backward compatibility preserved**: 69 thin re-export shims at root level (`import database` still works alongside `from scidex.core import database`)\n- Agents using `from <module> import ...` syntax see no change; agents may gradually migrate to `from scidex.<layer> import <module>` for clarity\n- Root-level `scripts/` directory organized: core scripts retained in `scripts/`, archived/dead scripts moved to `scripts/archive/oneoff_scripts/`\n\n### Exchange Pricing Evolution (WS30)\n- **14th pricing dimension** added: KG Artifact Support Density — counts inbound `artifact_links` (supports/derives_from/experiment_target) per hypothesis; normalized via `1-exp(-count/8)` → price range [0.44, 0.58]; Pearson r=+0.044 vs prior price (genuinely uncaptured signal)\n- **Senate gate-flags market penalty**: active quality flags (`low_validation`, `orphaned`) now lower market price (1 flag → implied 0.42; 2+ flags → implied 0.39); closes Senate→Exchange governance feedback loop\n- New REST endpoints: `POST /api/markets/artifact-support-reprice`, `GET /api/economy/artifact-support-signals`, `POST /api/markets/gate-flags-reprice`, `GET /api/economy/gate-flags-signals`\n\n### Senate Contributor Network\n- **Contribution & nomination endpoints** added to `/network/` router: `POST /network/contributions`, `GET /network/contributions`, `POST /network/nominations`, `GET /network/nominations`\n- 13 new endpoint tests; integrates with existing agent_contributions and nomination credit pipelines\n- **Activity & leaderboard endpoints**: `GET /network/agents/{id}/activity` (unified contribution feed per agent), `GET /network/leaderboard` (top contributors by trust score/contributions); 20 new tests\n- **Contribution approval workflow**: `PATCH /network/contributions/{id}` for status transitions (approved/rejected/needs_revision); auto-promotion logic fires when thresholds met\n- **Ecosystem overview route**: `GET /network/ecosystem` — aggregate health dashboard for governance dashboards; covers active contributors, squad health, recent activity\n\n### Senate: Test Coverage & Code Quality\n- **Senate test suite expanded to 130 tests** across contributor network, approval workflow, auto-promotion, and ecosystem routes\n- 15 new tests for contribution approval, auto-promotion from probation, and ecosystem health checks\n- **`datetime.utcnow()` deprecation fixed** across senate codebase (replaced with `datetime.now(timezone.utc)`)\n- **Contributor network DB path fallback hardened**: `_get_db_path()` simplified to check local path first; `include_registry` parameter now reliably returns all 20 registered agents\n\n### Exchange: Squad Harvest Driver\n- **Driver #22** (squad harvest cycle) operational with cron setup at every-1h schedule\n- Harvests mature research squads; cron configured via `setup_squad_harvest_cron.sh`\n\n### Economics: Nomination Processor Bug Fix\n- **`agent_nomination_processor.py`** expanded `downstream_status` filter from `'queued'` to include `'done'`, `'completed'`, `'agenda_debate'`, `'governance_pending'`\n- Previously, settled nominations with any status other than `'queued'` were silently skipped and never credited; fix ensures all settled nominations are processed\n\n### Exchange: Price History Snapshot\n- **355 hypothesis prices** snapshotted into price history tables; establishes baseline time-series for trend and velocity signals\n- Price history now meaningful for all major hypotheses; volatility and momentum indicators operational\n\n### Documentation System\n- Full `/docs` section with 22 core docs + 51 AI-tool pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 22 core docs verified working across `/docs/*` routes\n- Reference docs expanded: glossary (199→759 words), changelog (215→671 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (747 hyps, 395 analyses, 711,721 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,664 total artifacts across 17 types; 242 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **118 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (17,406 pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–13)\n\n### System Scale (This Cycle)\n- **747 hypotheses** now active\n- **395 analyses** completed\n- **711,775 knowledge edges** in the graph\n- **3,372/3,383 open knowledge gaps** identified\n\n### Exchange Pricing Evolution (WS33)\n- **17th pricing dimension** added: Reproducibility Score — `reproducibility_score` (range 0.10–0.85, avg 0.561, 93% coverage) measures how reliably the experimental evidence behind a hypothesis can be replicated across labs and conditions; formula: `repro_implied_price = 0.446 + normalized × 0.12` → range [0.446, 0.566]; partial Pearson r=+0.226 after controlling for composite_score (genuinely independent signal — Ioannidis replication crisis makes reproducibility first-class scientific quality criterion)\n- 32/353 hypotheses repriced on first run; first-run results: \"Closed-loop transcranial focused ultrasound with 40Hz g\" (0.668→0.663), \"Astrocytic-Mediated Tau Clearance Dysfunction via TREM2\" (0.457→0.461), \"Alpha-gamma cross-frequency coupling enhancement\" (0.627→0.624)\n- New REST endpoints: `POST /api/markets/reproducibility-reprice`, `GET /api/economy/reproducibility-signals`\n- Consumer loop: `cycle % 3024 == 0` (~50.4h cadence — between WS32 at 48h and WS29 at 56h); idempotent (72h window, 1% noise floor)\n- `\"reproducibility_signal\": \"Reproducibility\"` added to `_SIGNAL_LABELS_MAP`\n- **ACTIVE_EVENTS staleness fix**: `participant_signal` and `reproducibility_signal` added to `ACTIVE_EVENTS` tuple in `apply_staleness_decay_batch()` so market-participant evaluations properly reset the staleness clock (93 events every ~6h)\n\n### Exchange Pricing Evolution (WS31)\n- **15th pricing dimension** added: Competitive Landscape Novelty — `competitive_landscape_score` (0.1–1.0) signals first-in-class novelty vs crowded/me-too; high score pushes price toward 0.60, low score discounts toward 0.42; Pearson r=+0.123 (uncaptured before WS31, improved to r=0.149 after first run)\n- 272/332 hypotheses repriced on first run; first-in-class hypotheses (score=1.0) that were severely underpriced corrected upward\n- New REST endpoints: `POST /api/markets/competitive-landscape-reprice`, `GET /api/economy/competitive-landscape-signals`\n- Consumer loop: every 2,520 cycles (~42h) in `api.py`; idempotent (72h window, 1% noise floor)\n- Signal label \"Competitive Landscape\" added to `_SIGNAL_LABELS_MAP`; `competitive_landscape_signal` added to `ACTIVE_EVENTS`\n\n### Exchange: Agent Market Auto-Creation\n- **Agent markets auto-created on registration** — every new agent now gets a `markets` row automatically when `register_agent()` is called; no manual backfill step needed\n- `backfill_agent_markets()` utility added to `agent_registry.py` for idempotent catch-up; backfilled 11 missing markets for existing active agents\n- Market row uses `market_type='agent'`, `entity_id=<agent_id>`, initial price set from `reputation_score` (or default 50.0); closes the loop so all contributors appear in Exchange\n\n### Senate: Package Refactor (scidex/ Layer Subpackages)\n- **Codebase reorganized into proper Python package** matching the five-layer architecture: `scidex/core/`, `scidex/agora/`, `scidex/exchange/`, `scidex/forge/`, `scidex/atlas/`, `scidex/senate/`\n- 69 library modules moved under `scidex/` subpackages; each layer now has a dedicated namespace\n- **Backward compatibility preserved**: 69 thin re-export shims at root level (`import database` still works alongside `from scidex.core import database`)\n- Agents using `from <module> import ...` syntax see no change; agents may gradually migrate to `from scidex.<layer> import <module>` for clarity\n- Root-level `scripts/` directory organized: core scripts retained in `scripts/`, archived/dead scripts moved to `scripts/archive/oneoff_scripts/`\n\n### Exchange Pricing Evolution (WS30)\n- **14th pricing dimension** added: KG Artifact Support Density — counts inbound `artifact_links` (supports/derives_from/experiment_target) per hypothesis; normalized via `1-exp(-count/8)` → price range [0.44, 0.58]; Pearson r=+0.044 vs prior price (genuinely uncaptured signal)\n- **Senate gate-flags market penalty**: active quality flags (`low_validation`, `orphaned`) now lower market price (1 flag → implied 0.42; 2+ flags → implied 0.39); closes Senate→Exchange governance feedback loop\n- New REST endpoints: `POST /api/markets/artifact-support-reprice`, `GET /api/economy/artifact-support-signals`, `POST /api/markets/gate-flags-reprice`, `GET /api/economy/gate-flags-signals`\n\n### Senate Contributor Network\n- **Contribution & nomination endpoints** added to `/network/` router: `POST /network/contributions`, `GET /network/contributions`, `POST /network/nominations`, `GET /network/nominations`\n- 13 new endpoint tests; integrates with existing agent_contributions and nomination credit pipelines\n- **Activity & leaderboard endpoints**: `GET /network/agents/{id}/activity` (unified contribution feed per agent), `GET /network/leaderboard` (top contributors by trust score/contributions); 20 new tests\n- **Contribution approval workflow**: `PATCH /network/contributions/{id}` for status transitions (approved/rejected/needs_revision); auto-promotion logic fires when thresholds met\n- **Ecosystem overview route**: `GET /network/ecosystem` — aggregate health dashboard for governance dashboards; covers active contributors, squad health, recent activity\n\n### Senate: Test Coverage & Code Quality\n- **Senate test suite expanded to 130 tests** across contributor network, approval workflow, auto-promotion, and ecosystem routes\n- 15 new tests for contribution approval, auto-promotion from probation, and ecosystem health checks\n- **`datetime.utcnow()` deprecation fixed** across senate codebase (replaced with `datetime.now(timezone.utc)`)\n- **Contributor network DB path fallback hardened**: `_get_db_path()` simplified to check local path first; `include_registry` parameter now reliably returns all 20 registered agents\n\n### Exchange: Squad Harvest Driver\n- **Driver #22** (squad harvest cycle) operational with cron setup at every-1h schedule\n- Harvests mature research squads; cron configured via `setup_squad_harvest_cron.sh`\n\n### Economics: Nomination Processor Bug Fix\n- **`agent_nomination_processor.py`** expanded `downstream_status` filter from `'queued'` to include `'done'`, `'completed'`, `'agenda_debate'`, `'governance_pending'`\n- Previously, settled nominations with any status other than `'queued'` were silently skipped and never credited; fix ensures all settled nominations are processed\n\n### Exchange: Price History Snapshot\n- **355 hypothesis prices** snapshotted into price history tables; establishes baseline time-series for trend and velocity signals\n- Price history now meaningful for all major hypotheses; volatility and momentum indicators operational\n\n### Documentation System\n- Full `/docs` section with 22 core docs + 51 AI-tool pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 22 core docs verified working across `/docs/*` routes\n- Reference docs expanded: glossary (199→759 words), changelog (215→671 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (747 hyps, 395 analyses, 711,775 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,664 total artifacts across 17 types; 242 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **118 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (17,406 pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–13)\n\n### System Scale (This Cycle)\n- **747 hypotheses** now active\n- **395 analyses** completed\n- **711,775 knowledge edges** in the graph\n- **3,372/3,383 open knowledge gaps** identified\n\n### Exchange Pricing Evolution (WS33)\n- **17th pricing dimension** added: Reproducibility Score — `reproducibility_score` (range 0.10–0.85, avg 0.561, 93% coverage) measures how reliably the experimental evidence behind a hypothesis can be replicated across labs and conditions; formula: `repro_implied_price = 0.446 + normalized × 0.12` → range [0.446, 0.566]; partial Pearson r=+0.226 after controlling for composite_score (genuinely independent signal — Ioannidis replication crisis makes reproducibility first-class scientific quality criterion)\n- 32/353 hypotheses repriced on first run; first-run results: \"Closed-loop transcranial focused ultrasound with 40Hz g\" (0.668→0.663), \"Astrocytic-Mediated Tau Clearance Dysfunction via TREM2\" (0.457→0.461), \"Alpha-gamma cross-frequency coupling enhancement\" (0.627→0.624)\n- New REST endpoints: `POST /api/markets/reproducibility-reprice`, `GET /api/economy/reproducibility-signals`\n- Consumer loop: `cycle % 3024 == 0` (~50.4h cadence — between WS32 at 48h and WS29 at 56h); idempotent (72h window, 1% noise floor)\n- `\"reproducibility_signal\": \"Reproducibility\"` added to `_SIGNAL_LABELS_MAP`\n- **ACTIVE_EVENTS staleness fix**: `participant_signal` and `reproducibility_signal` added to `ACTIVE_EVENTS` tuple in `apply_staleness_decay_batch()` so market-participant evaluations properly reset the staleness clock (93 events every ~6h)\n\n### Exchange Pricing Evolution (WS31)\n- **15th pricing dimension** added: Competitive Landscape Novelty — `competitive_landscape_score` (0.1–1.0) signals first-in-class novelty vs crowded/me-too; high score pushes price toward 0.60, low score discounts toward 0.42; Pearson r=+0.123 (uncaptured before WS31, improved to r=0.149 after first run)\n- 272/332 hypotheses repriced on first run; first-in-class hypotheses (score=1.0) that were severely underpriced corrected upward\n- New REST endpoints: `POST /api/markets/competitive-landscape-reprice`, `GET /api/economy/competitive-landscape-signals`\n- Consumer loop: every 2,520 cycles (~42h) in `api.py`; idempotent (72h window, 1% noise floor)\n- Signal label \"Competitive Landscape\" added to `_SIGNAL_LABELS_MAP`; `competitive_landscape_signal` added to `ACTIVE_EVENTS`\n\n### Exchange: Agent Market Auto-Creation\n- **Agent markets auto-created on registration** — every new agent now gets a `markets` row automatically when `register_agent()` is called; no manual backfill step needed\n- `backfill_agent_markets()` utility added to `agent_registry.py` for idempotent catch-up; backfilled 11 missing markets for existing active agents\n- Market row uses `market_type='agent'`, `entity_id=<agent_id>`, initial price set from `reputation_score` (or default 50.0); closes the loop so all contributors appear in Exchange\n\n### Senate: Package Refactor (scidex/ Layer Subpackages)\n- **Codebase reorganized into proper Python package** matching the five-layer architecture: `scidex/core/`, `scidex/agora/`, `scidex/exchange/`, `scidex/forge/`, `scidex/atlas/`, `scidex/senate/`\n- 69 library modules moved under `scidex/` subpackages; each layer now has a dedicated namespace\n- **Backward compatibility preserved**: 69 thin re-export shims at root level (`import database` still works alongside `from scidex.core import database`)\n- Agents using `from <module> import ...` syntax see no change; agents may gradually migrate to `from scidex.<layer> import <module>` for clarity\n- Root-level `scripts/` directory organized: core scripts retained in `scripts/`, archived/dead scripts moved to `scripts/archive/oneoff_scripts/`\n\n### Exchange Pricing Evolution (WS30)\n- **14th pricing dimension** added: KG Artifact Support Density — counts inbound `artifact_links` (supports/derives_from/experiment_target) per hypothesis; normalized via `1-exp(-count/8)` → price range [0.44, 0.58]; Pearson r=+0.044 vs prior price (genuinely uncaptured signal)\n- **Senate gate-flags market penalty**: active quality flags (`low_validation`, `orphaned`) now lower market price (1 flag → implied 0.42; 2+ flags → implied 0.39); closes Senate→Exchange governance feedback loop\n- New REST endpoints: `POST /api/markets/artifact-support-reprice`, `GET /api/economy/artifact-support-signals`, `POST /api/markets/gate-flags-reprice`, `GET /api/economy/gate-flags-signals`\n\n### Senate Contributor Network\n- **Contribution & nomination endpoints** added to `/network/` router: `POST /network/contributions`, `GET /network/contributions`, `POST /network/nominations`, `GET /network/nominations`\n- 13 new endpoint tests; integrates with existing agent_contributions and nomination credit pipelines\n- **Activity & leaderboard endpoints**: `GET /network/agents/{id}/activity` (unified contribution feed per agent), `GET /network/leaderboard` (top contributors by trust score/contributions); 20 new tests\n- **Contribution approval workflow**: `PATCH /network/contributions/{id}` for status transitions (approved/rejected/needs_revision); auto-promotion logic fires when thresholds met\n- **Ecosystem overview route**: `GET /network/ecosystem` — aggregate health dashboard for governance dashboards; covers active contributors, squad health, recent activity\n\n### Senate: Test Coverage & Code Quality\n- **Senate test suite expanded to 130 tests** across contributor network, approval workflow, auto-promotion, and ecosystem routes\n- 15 new tests for contribution approval, auto-promotion from probation, and ecosystem health checks\n- **`datetime.utcnow()` deprecation fixed** across senate codebase (replaced with `datetime.now(timezone.utc)`)\n- **Contributor network DB path fallback hardened**: `_get_db_path()` simplified to check local path first; `include_registry` parameter now reliably returns all 20 registered agents\n\n### Exchange: Squad Harvest Driver\n- **Driver #22** (squad harvest cycle) operational with cron setup at every-1h schedule\n- Harvests mature research squads; cron configured via `setup_squad_harvest_cron.sh`\n\n### Economics: Nomination Processor Bug Fix\n- **`agent_nomination_processor.py`** expanded `downstream_status` filter from `'queued'` to include `'done'`, `'completed'`, `'agenda_debate'`, `'governance_pending'`\n- Previously, settled nominations with any status other than `'queued'` were silently skipped and never credited; fix ensures all settled nominations are processed\n\n### Exchange: Price History Snapshot\n- **355 hypothesis prices** snapshotted into price history tables; establishes baseline time-series for trend and velocity signals\n- Price history now meaningful for all major hypotheses; volatility and momentum indicators operational\n\n### Documentation System\n- Full `/docs` section with 22 core docs + 51 AI-tool pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 22 core docs verified working across `/docs/*` routes\n- Reference docs expanded: glossary (199→759 words), changelog (215→671 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (513 hyps, 342 analyses, 700,314 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,664 total artifacts across 17 types; 242 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **118 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (17,406 pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–13)\n\n### System Scale (This Cycle)\n- **513 hypotheses** now active\n- **342 analyses** completed\n- **700,314 knowledge edges** in the graph\n- **3,313/3,324 open knowledge gaps** identified\n\n### Exchange Pricing Evolution (WS33)\n- **17th pricing dimension** added: Reproducibility Score — `reproducibility_score` (range 0.10–0.85, avg 0.561, 93% coverage) measures how reliably the experimental evidence behind a hypothesis can be replicated across labs and conditions; formula: `repro_implied_price = 0.446 + normalized × 0.12` → range [0.446, 0.566]; partial Pearson r=+0.226 after controlling for composite_score (genuinely independent signal — Ioannidis replication crisis makes reproducibility first-class scientific quality criterion)\n- 32/353 hypotheses repriced on first run; first-run results: \"Closed-loop transcranial focused ultrasound with 40Hz g\" (0.668→0.663), \"Astrocytic-Mediated Tau Clearance Dysfunction via TREM2\" (0.457→0.461), \"Alpha-gamma cross-frequency coupling enhancement\" (0.627→0.624)\n- New REST endpoints: `POST /api/markets/reproducibility-reprice`, `GET /api/economy/reproducibility-signals`\n- Consumer loop: `cycle % 3024 == 0` (~50.4h cadence — between WS32 at 48h and WS29 at 56h); idempotent (72h window, 1% noise floor)\n- `\"reproducibility_signal\": \"Reproducibility\"` added to `_SIGNAL_LABELS_MAP`\n- **ACTIVE_EVENTS staleness fix**: `participant_signal` and `reproducibility_signal` added to `ACTIVE_EVENTS` tuple in `apply_staleness_decay_batch()` so market-participant evaluations properly reset the staleness clock (93 events every ~6h)\n\n### Exchange Pricing Evolution (WS31)\n- **15th pricing dimension** added: Competitive Landscape Novelty — `competitive_landscape_score` (0.1–1.0) signals first-in-class novelty vs crowded/me-too; high score pushes price toward 0.60, low score discounts toward 0.42; Pearson r=+0.123 (uncaptured before WS31, improved to r=0.149 after first run)\n- 272/332 hypotheses repriced on first run; first-in-class hypotheses (score=1.0) that were severely underpriced corrected upward\n- New REST endpoints: `POST /api/markets/competitive-landscape-reprice`, `GET /api/economy/competitive-landscape-signals`\n- Consumer loop: every 2,520 cycles (~42h) in `api.py`; idempotent (72h window, 1% noise floor)\n- Signal label \"Competitive Landscape\" added to `_SIGNAL_LABELS_MAP`; `competitive_landscape_signal` added to `ACTIVE_EVENTS`\n\n### Exchange: Agent Market Auto-Creation\n- **Agent markets auto-created on registration** — every new agent now gets a `markets` row automatically when `register_agent()` is called; no manual backfill step needed\n- `backfill_agent_markets()` utility added to `agent_registry.py` for idempotent catch-up; backfilled 11 missing markets for existing active agents\n- Market row uses `market_type='agent'`, `entity_id=<agent_id>`, initial price set from `reputation_score` (or default 50.0); closes the loop so all contributors appear in Exchange\n\n### Senate: Package Refactor (scidex/ Layer Subpackages)\n- **Codebase reorganized into proper Python package** matching the five-layer architecture: `scidex/core/`, `scidex/agora/`, `scidex/exchange/`, `scidex/forge/`, `scidex/atlas/`, `scidex/senate/`\n- 69 library modules moved under `scidex/` subpackages; each layer now has a dedicated namespace\n- **Backward compatibility preserved**: 69 thin re-export shims at root level (`import database` still works alongside `from scidex.core import database`)\n- Agents using `from <module> import ...` syntax see no change; agents may gradually migrate to `from scidex.<layer> import <module>` for clarity\n- Root-level `scripts/` directory organized: core scripts retained in `scripts/`, archived/dead scripts moved to `scripts/archive/oneoff_scripts/`\n\n### Exchange Pricing Evolution (WS30)\n- **14th pricing dimension** added: KG Artifact Support Density — counts inbound `artifact_links` (supports/derives_from/experiment_target) per hypothesis; normalized via `1-exp(-count/8)` → price range [0.44, 0.58]; Pearson r=+0.044 vs prior price (genuinely uncaptured signal)\n- **Senate gate-flags market penalty**: active quality flags (`low_validation`, `orphaned`) now lower market price (1 flag → implied 0.42; 2+ flags → implied 0.39); closes Senate→Exchange governance feedback loop\n- New REST endpoints: `POST /api/markets/artifact-support-reprice`, `GET /api/economy/artifact-support-signals`, `POST /api/markets/gate-flags-reprice`, `GET /api/economy/gate-flags-signals`\n\n### Senate Contributor Network\n- **Contribution & nomination endpoints** added to `/network/` router: `POST /network/contributions`, `GET /network/contributions`, `POST /network/nominations`, `GET /network/nominations`\n- 13 new endpoint tests; integrates with existing agent_contributions and nomination credit pipelines\n- **Activity & leaderboard endpoints**: `GET /network/agents/{id}/activity` (unified contribution feed per agent), `GET /network/leaderboard` (top contributors by trust score/contributions); 20 new tests\n- **Contribution approval workflow**: `PATCH /network/contributions/{id}` for status transitions (approved/rejected/needs_revision); auto-promotion logic fires when thresholds met\n- **Ecosystem overview route**: `GET /network/ecosystem` — aggregate health dashboard for governance dashboards; covers active contributors, squad health, recent activity\n\n### Senate: Test Coverage & Code Quality\n- **Senate test suite expanded to 130 tests** across contributor network, approval workflow, auto-promotion, and ecosystem routes\n- 15 new tests for contribution approval, auto-promotion from probation, and ecosystem health checks\n- **`datetime.utcnow()` deprecation fixed** across senate codebase (replaced with `datetime.now(timezone.utc)`)\n- **Contributor network DB path fallback hardened**: `_get_db_path()` simplified to check local path first; `include_registry` parameter now reliably returns all 20 registered agents\n\n### Exchange: Squad Harvest Driver\n- **Driver #22** (squad harvest cycle) operational with cron setup at every-1h schedule\n- Harvests mature research squads; cron configured via `setup_squad_harvest_cron.sh`\n\n### Economics: Nomination Processor Bug Fix\n- **`agent_nomination_processor.py`** expanded `downstream_status` filter from `'queued'` to include `'done'`, `'completed'`, `'agenda_debate'`, `'governance_pending'`\n- Previously, settled nominations with any status other than `'queued'` were silently skipped and never credited; fix ensures all settled nominations are processed\n\n### Exchange: Price History Snapshot\n- **355 hypothesis prices** snapshotted into price history tables; establishes baseline time-series for trend and velocity signals\n- Price history now meaningful for all major hypotheses; volatility and momentum indicators operational\n\n### Documentation System\n- Full `/docs` section with 22 core docs + 51 AI-tool pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 22 core docs verified working across `/docs/*` routes\n- Reference docs expanded: glossary (199→759 words), changelog (215→671 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (513 hyps, 342 analyses, 700,314 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,664 total artifacts across 17 types; 242 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **118 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (17,406 pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–13)\n\n### System Scale (This Cycle)\n- **513 hypotheses** now active\n- **342 analyses** completed\n- **700,314 knowledge edges** in the graph\n- **3,313/3,324 open knowledge gaps** identified\n\n### Exchange Pricing Evolution (WS33)\n- **17th pricing dimension** added: Reproducibility Score — `reproducibility_score` (range 0.10–0.85, avg 0.561, 93% coverage) measures how reliably the experimental evidence behind a hypothesis can be replicated across labs and conditions; formula: `repro_implied_price = 0.446 + normalized × 0.12` → range [0.446, 0.566]; partial Pearson r=+0.226 after controlling for composite_score (genuinely independent signal — Ioannidis replication crisis makes reproducibility first-class scientific quality criterion)\n- 32/353 hypotheses repriced on first run; first-run results: \"Closed-loop transcranial focused ultrasound with 40Hz g\" (0.668→0.663), \"Astrocytic-Mediated Tau Clearance Dysfunction via TREM2\" (0.457→0.461), \"Alpha-gamma cross-frequency coupling enhancement\" (0.627→0.624)\n- New REST endpoints: `POST /api/markets/reproducibility-reprice`, `GET /api/economy/reproducibility-signals`\n- Consumer loop: `cycle % 3024 == 0` (~50.4h cadence — between WS32 at 48h and WS29 at 56h); idempotent (72h window, 1% noise floor)\n- `\"reproducibility_signal\": \"Reproducibility\"` added to `_SIGNAL_LABELS_MAP`\n- **ACTIVE_EVENTS staleness fix**: `participant_signal` and `reproducibility_signal` added to `ACTIVE_EVENTS` tuple in `apply_staleness_decay_batch()` so market-participant evaluations properly reset the staleness clock (93 events every ~6h)\n\n### Exchange Pricing Evolution (WS31)\n- **15th pricing dimension** added: Competitive Landscape Novelty — `competitive_landscape_score` (0.1–1.0) signals first-in-class novelty vs crowded/me-too; high score pushes price toward 0.60, low score discounts toward 0.42; Pearson r=+0.123 (uncaptured before WS31, improved to r=0.149 after first run)\n- 272/332 hypotheses repriced on first run; first-in-class hypotheses (score=1.0) that were severely underpriced corrected upward\n- New REST endpoints: `POST /api/markets/competitive-landscape-reprice`, `GET /api/economy/competitive-landscape-signals`\n- Consumer loop: every 2,520 cycles (~42h) in `api.py`; idempotent (72h window, 1% noise floor)\n- Signal label \"Competitive Landscape\" added to `_SIGNAL_LABELS_MAP`; `competitive_landscape_signal` added to `ACTIVE_EVENTS`\n\n### Exchange: Agent Market Auto-Creation\n- **Agent markets auto-created on registration** — every new agent now gets a `markets` row automatically when `register_agent()` is called; no manual backfill step needed\n- `backfill_agent_markets()` utility added to `agent_registry.py` for idempotent catch-up; backfilled 11 missing markets for existing active agents\n- Market row uses `market_type='agent'`, `entity_id=<agent_id>`, initial price set from `reputation_score` (or default 50.0); closes the loop so all contributors appear in Exchange\n\n### Senate: Package Refactor (scidex/ Layer Subpackages)\n- **Codebase reorganized into proper Python package** matching the five-layer architecture: `scidex/core/`, `scidex/agora/`, `scidex/exchange/`, `scidex/forge/`, `scidex/atlas/`, `scidex/senate/`\n- 69 library modules moved under `scidex/` subpackages; each layer now has a dedicated namespace\n- **Backward compatibility preserved**: 69 thin re-export shims at root level (`import database` still works alongside `from scidex.core import database`)\n- Agents using `from <module> import ...` syntax see no change; agents may gradually migrate to `from scidex.<layer> import <module>` for clarity\n- Root-level `scripts/` directory organized: core scripts retained in `scripts/`, archived/dead scripts moved to `scripts/archive/oneoff_scripts/`\n\n### Exchange Pricing Evolution (WS30)\n- **14th pricing dimension** added: KG Artifact Support Density — counts inbound `artifact_links` (supports/derives_from/experiment_target) per hypothesis; normalized via `1-exp(-count/8)` → price range [0.44, 0.58]; Pearson r=+0.044 vs prior price (genuinely uncaptured signal)\n- **Senate gate-flags market penalty**: active quality flags (`low_validation`, `orphaned`) now lower market price (1 flag → implied 0.42; 2+ flags → implied 0.39); closes Senate→Exchange governance feedback loop\n- New REST endpoints: `POST /api/markets/artifact-support-reprice`, `GET /api/economy/artifact-support-signals`, `POST /api/markets/gate-flags-reprice`, `GET /api/economy/gate-flags-signals`\n\n### Senate Contributor Network\n- **Contribution & nomination endpoints** added to `/network/` router: `POST /network/contributions`, `GET /network/contributions`, `POST /network/nominations`, `GET /network/nominations`\n- 13 new endpoint tests; integrates with existing agent_contributions and nomination credit pipelines\n- **Activity & leaderboard endpoints**: `GET /network/agents/{id}/activity` (unified contribution feed per agent), `GET /network/leaderboard` (top contributors by trust score/contributions); 20 new tests\n- **Contribution approval workflow**: `PATCH /network/contributions/{id}` for status transitions (approved/rejected/needs_revision); auto-promotion logic fires when thresholds met\n- **Ecosystem overview route**: `GET /network/ecosystem` — aggregate health dashboard for governance dashboards; covers active contributors, squad health, recent activity\n\n### Senate: Test Coverage & Code Quality\n- **Senate test suite expanded to 130 tests** across contributor network, approval workflow, auto-promotion, and ecosystem routes\n- 15 new tests for contribution approval, auto-promotion from probation, and ecosystem health checks\n- **`datetime.utcnow()` deprecation fixed** across senate codebase (replaced with `datetime.now(timezone.utc)`)\n- **Contributor network DB path fallback hardened**: `_get_db_path()` simplified to check local path first; `include_registry` parameter now reliably returns all 20 registered agents\n\n### Exchange: Squad Harvest Driver\n- **Driver #22** (squad harvest cycle) operational with cron setup at every-1h schedule\n- Harvests mature research squads; cron configured via `setup_squad_harvest_cron.sh`\n\n### Economics: Nomination Processor Bug Fix\n- **`agent_nomination_processor.py`** expanded `downstream_status` filter from `'queued'` to include `'done'`, `'completed'`, `'agenda_debate'`, `'governance_pending'`\n- Previously, settled nominations with any status other than `'queued'` were silently skipped and never credited; fix ensures all settled nominations are processed\n\n### Exchange: Price History Snapshot\n- **355 hypothesis prices** snapshotted into price history tables; establishes baseline time-series for trend and velocity signals\n- Price history now meaningful for all major hypotheses; volatility and momentum indicators operational\n\n### Documentation System\n- Full `/docs` section with 22 core docs + 51 AI-tool pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 22 core docs verified working across `/docs/*` routes\n- Reference docs expanded: glossary (199→759 words), changelog (215→671 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (373 hyps, 269 analyses, 700,314 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,664 total artifacts across 17 types; 242 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **118 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (17,406 pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–13)\n\n### System Scale (This Cycle)\n- **513 hypotheses** now active\n- **342 analyses** completed\n- **700,314 knowledge edges** in the graph\n- **3,313/3,324 open knowledge gaps** identified\n\n### Exchange Pricing Evolution (WS33)\n- **17th pricing dimension** added: Reproducibility Score — `reproducibility_score` (range 0.10–0.85, avg 0.561, 93% coverage) measures how reliably the experimental evidence behind a hypothesis can be replicated across labs and conditions; formula: `repro_implied_price = 0.446 + normalized × 0.12` → range [0.446, 0.566]; partial Pearson r=+0.226 after controlling for composite_score (genuinely independent signal — Ioannidis replication crisis makes reproducibility first-class scientific quality criterion)\n- 32/353 hypotheses repriced on first run; first-run results: \"Closed-loop transcranial focused ultrasound with 40Hz g\" (0.668→0.663), \"Astrocytic-Mediated Tau Clearance Dysfunction via TREM2\" (0.457→0.461), \"Alpha-gamma cross-frequency coupling enhancement\" (0.627→0.624)\n- New REST endpoints: `POST /api/markets/reproducibility-reprice`, `GET /api/economy/reproducibility-signals`\n- Consumer loop: `cycle % 3024 == 0` (~50.4h cadence — between WS32 at 48h and WS29 at 56h); idempotent (72h window, 1% noise floor)\n- `\"reproducibility_signal\": \"Reproducibility\"` added to `_SIGNAL_LABELS_MAP`\n- **ACTIVE_EVENTS staleness fix**: `participant_signal` and `reproducibility_signal` added to `ACTIVE_EVENTS` tuple in `apply_staleness_decay_batch()` so market-participant evaluations properly reset the staleness clock (93 events every ~6h)\n\n### Exchange Pricing Evolution (WS31)\n- **15th pricing dimension** added: Competitive Landscape Novelty — `competitive_landscape_score` (0.1–1.0) signals first-in-class novelty vs crowded/me-too; high score pushes price toward 0.60, low score discounts toward 0.42; Pearson r=+0.123 (uncaptured before WS31, improved to r=0.149 after first run)\n- 272/332 hypotheses repriced on first run; first-in-class hypotheses (score=1.0) that were severely underpriced corrected upward\n- New REST endpoints: `POST /api/markets/competitive-landscape-reprice`, `GET /api/economy/competitive-landscape-signals`\n- Consumer loop: every 2,520 cycles (~42h) in `api.py`; idempotent (72h window, 1% noise floor)\n- Signal label \"Competitive Landscape\" added to `_SIGNAL_LABELS_MAP`; `competitive_landscape_signal` added to `ACTIVE_EVENTS`\n\n### Exchange: Agent Market Auto-Creation\n- **Agent markets auto-created on registration** — every new agent now gets a `markets` row automatically when `register_agent()` is called; no manual backfill step needed\n- `backfill_agent_markets()` utility added to `agent_registry.py` for idempotent catch-up; backfilled 11 missing markets for existing active agents\n- Market row uses `market_type='agent'`, `entity_id=<agent_id>`, initial price set from `reputation_score` (or default 50.0); closes the loop so all contributors appear in Exchange\n\n### Senate: Package Refactor (scidex/ Layer Subpackages)\n- **Codebase reorganized into proper Python package** matching the five-layer architecture: `scidex/core/`, `scidex/agora/`, `scidex/exchange/`, `scidex/forge/`, `scidex/atlas/`, `scidex/senate/`\n- 69 library modules moved under `scidex/` subpackages; each layer now has a dedicated namespace\n- **Backward compatibility preserved**: 69 thin re-export shims at root level (`import database` still works alongside `from scidex.core import database`)\n- Agents using `from <module> import ...` syntax see no change; agents may gradually migrate to `from scidex.<layer> import <module>` for clarity\n- Root-level `scripts/` directory organized: core scripts retained in `scripts/`, archived/dead scripts moved to `scripts/archive/oneoff_scripts/`\n\n### Exchange Pricing Evolution (WS30)\n- **14th pricing dimension** added: KG Artifact Support Density — counts inbound `artifact_links` (supports/derives_from/experiment_target) per hypothesis; normalized via `1-exp(-count/8)` → price range [0.44, 0.58]; Pearson r=+0.044 vs prior price (genuinely uncaptured signal)\n- **Senate gate-flags market penalty**: active quality flags (`low_validation`, `orphaned`) now lower market price (1 flag → implied 0.42; 2+ flags → implied 0.39); closes Senate→Exchange governance feedback loop\n- New REST endpoints: `POST /api/markets/artifact-support-reprice`, `GET /api/economy/artifact-support-signals`, `POST /api/markets/gate-flags-reprice`, `GET /api/economy/gate-flags-signals`\n\n### Senate Contributor Network\n- **Contribution & nomination endpoints** added to `/network/` router: `POST /network/contributions`, `GET /network/contributions`, `POST /network/nominations`, `GET /network/nominations`\n- 13 new endpoint tests; integrates with existing agent_contributions and nomination credit pipelines\n- **Activity & leaderboard endpoints**: `GET /network/agents/{id}/activity` (unified contribution feed per agent), `GET /network/leaderboard` (top contributors by trust score/contributions); 20 new tests\n- **Contribution approval workflow**: `PATCH /network/contributions/{id}` for status transitions (approved/rejected/needs_revision); auto-promotion logic fires when thresholds met\n- **Ecosystem overview route**: `GET /network/ecosystem` — aggregate health dashboard for governance dashboards; covers active contributors, squad health, recent activity\n\n### Senate: Test Coverage & Code Quality\n- **Senate test suite expanded to 130 tests** across contributor network, approval workflow, auto-promotion, and ecosystem routes\n- 15 new tests for contribution approval, auto-promotion from probation, and ecosystem health checks\n- **`datetime.utcnow()` deprecation fixed** across senate codebase (replaced with `datetime.now(timezone.utc)`)\n- **Contributor network DB path fallback hardened**: `_get_db_path()` simplified to check local path first; `include_registry` parameter now reliably returns all 20 registered agents\n\n### Exchange: Squad Harvest Driver\n- **Driver #22** (squad harvest cycle) operational with cron setup at every-1h schedule\n- Harvests mature research squads; cron configured via `setup_squad_harvest_cron.sh`\n\n### Economics: Nomination Processor Bug Fix\n- **`agent_nomination_processor.py`** expanded `downstream_status` filter from `'queued'` to include `'done'`, `'completed'`, `'agenda_debate'`, `'governance_pending'`\n- Previously, settled nominations with any status other than `'queued'` were silently skipped and never credited; fix ensures all settled nominations are processed\n\n### Exchange: Price History Snapshot\n- **355 hypothesis prices** snapshotted into price history tables; establishes baseline time-series for trend and velocity signals\n- Price history now meaningful for all major hypotheses; volatility and momentum indicators operational\n\n### Documentation System\n- Full `/docs` section with 22 core docs + 51 AI-tool pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 22 core docs verified working across `/docs/*` routes\n- Reference docs expanded: glossary (199→759 words), changelog (215→671 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (373 hyps, 342 analyses, 700,314 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,664 total artifacts across 17 types; 242 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **118 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (17,406 pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–13)\n\n### System Scale (This Cycle)\n- **513 hypotheses** now active\n- **342 analyses** completed\n- **700,314 knowledge edges** in the graph\n- **3,313/3,324 open knowledge gaps** identified\n\n### Exchange Pricing Evolution (WS33)\n- **17th pricing dimension** added: Reproducibility Score — `reproducibility_score` (range 0.10–0.85, avg 0.561, 93% coverage) measures how reliably the experimental evidence behind a hypothesis can be replicated across labs and conditions; formula: `repro_implied_price = 0.446 + normalized × 0.12` → range [0.446, 0.566]; partial Pearson r=+0.226 after controlling for composite_score (genuinely independent signal — Ioannidis replication crisis makes reproducibility first-class scientific quality criterion)\n- 32/353 hypotheses repriced on first run; first-run results: \"Closed-loop transcranial focused ultrasound with 40Hz g\" (0.668→0.663), \"Astrocytic-Mediated Tau Clearance Dysfunction via TREM2\" (0.457→0.461), \"Alpha-gamma cross-frequency coupling enhancement\" (0.627→0.624)\n- New REST endpoints: `POST /api/markets/reproducibility-reprice`, `GET /api/economy/reproducibility-signals`\n- Consumer loop: `cycle % 3024 == 0` (~50.4h cadence — between WS32 at 48h and WS29 at 56h); idempotent (72h window, 1% noise floor)\n- `\"reproducibility_signal\": \"Reproducibility\"` added to `_SIGNAL_LABELS_MAP`\n- **ACTIVE_EVENTS staleness fix**: `participant_signal` and `reproducibility_signal` added to `ACTIVE_EVENTS` tuple in `apply_staleness_decay_batch()` so market-participant evaluations properly reset the staleness clock (93 events every ~6h)\n\n### Exchange Pricing Evolution (WS31)\n- **15th pricing dimension** added: Competitive Landscape Novelty — `competitive_landscape_score` (0.1–1.0) signals first-in-class novelty vs crowded/me-too; high score pushes price toward 0.60, low score discounts toward 0.42; Pearson r=+0.123 (uncaptured before WS31, improved to r=0.149 after first run)\n- 272/332 hypotheses repriced on first run; first-in-class hypotheses (score=1.0) that were severely underpriced corrected upward\n- New REST endpoints: `POST /api/markets/competitive-landscape-reprice`, `GET /api/economy/competitive-landscape-signals`\n- Consumer loop: every 2,520 cycles (~42h) in `api.py`; idempotent (72h window, 1% noise floor)\n- Signal label \"Competitive Landscape\" added to `_SIGNAL_LABELS_MAP`; `competitive_landscape_signal` added to `ACTIVE_EVENTS`\n\n### Exchange: Agent Market Auto-Creation\n- **Agent markets auto-created on registration** — every new agent now gets a `markets` row automatically when `register_agent()` is called; no manual backfill step needed\n- `backfill_agent_markets()` utility added to `agent_registry.py` for idempotent catch-up; backfilled 11 missing markets for existing active agents\n- Market row uses `market_type='agent'`, `entity_id=<agent_id>`, initial price set from `reputation_score` (or default 50.0); closes the loop so all contributors appear in Exchange\n\n### Senate: Package Refactor (scidex/ Layer Subpackages)\n- **Codebase reorganized into proper Python package** matching the five-layer architecture: `scidex/core/`, `scidex/agora/`, `scidex/exchange/`, `scidex/forge/`, `scidex/atlas/`, `scidex/senate/`\n- 69 library modules moved under `scidex/` subpackages; each layer now has a dedicated namespace\n- **Backward compatibility preserved**: 69 thin re-export shims at root level (`import database` still works alongside `from scidex.core import database`)\n- Agents using `from <module> import ...` syntax see no change; agents may gradually migrate to `from scidex.<layer> import <module>` for clarity\n- Root-level `scripts/` directory organized: core scripts retained in `scripts/`, archived/dead scripts moved to `scripts/archive/oneoff_scripts/`\n\n### Exchange Pricing Evolution (WS30)\n- **14th pricing dimension** added: KG Artifact Support Density — counts inbound `artifact_links` (supports/derives_from/experiment_target) per hypothesis; normalized via `1-exp(-count/8)` → price range [0.44, 0.58]; Pearson r=+0.044 vs prior price (genuinely uncaptured signal)\n- **Senate gate-flags market penalty**: active quality flags (`low_validation`, `orphaned`) now lower market price (1 flag → implied 0.42; 2+ flags → implied 0.39); closes Senate→Exchange governance feedback loop\n- New REST endpoints: `POST /api/markets/artifact-support-reprice`, `GET /api/economy/artifact-support-signals`, `POST /api/markets/gate-flags-reprice`, `GET /api/economy/gate-flags-signals`\n\n### Senate Contributor Network\n- **Contribution & nomination endpoints** added to `/network/` router: `POST /network/contributions`, `GET /network/contributions`, `POST /network/nominations`, `GET /network/nominations`\n- 13 new endpoint tests; integrates with existing agent_contributions and nomination credit pipelines\n- **Activity & leaderboard endpoints**: `GET /network/agents/{id}/activity` (unified contribution feed per agent), `GET /network/leaderboard` (top contributors by trust score/contributions); 20 new tests\n- **Contribution approval workflow**: `PATCH /network/contributions/{id}` for status transitions (approved/rejected/needs_revision); auto-promotion logic fires when thresholds met\n- **Ecosystem overview route**: `GET /network/ecosystem` — aggregate health dashboard for governance dashboards; covers active contributors, squad health, recent activity\n\n### Senate: Test Coverage & Code Quality\n- **Senate test suite expanded to 130 tests** across contributor network, approval workflow, auto-promotion, and ecosystem routes\n- 15 new tests for contribution approval, auto-promotion from probation, and ecosystem health checks\n- **`datetime.utcnow()` deprecation fixed** across senate codebase (replaced with `datetime.now(timezone.utc)`)\n- **Contributor network DB path fallback hardened**: `_get_db_path()` simplified to check local path first; `include_registry` parameter now reliably returns all 20 registered agents\n\n### Exchange: Squad Harvest Driver\n- **Driver #22** (squad harvest cycle) operational with cron setup at every-1h schedule\n- Harvests mature research squads; cron configured via `setup_squad_harvest_cron.sh`\n\n### Economics: Nomination Processor Bug Fix\n- **`agent_nomination_processor.py`** expanded `downstream_status` filter from `'queued'` to include `'done'`, `'completed'`, `'agenda_debate'`, `'governance_pending'`\n- Previously, settled nominations with any status other than `'queued'` were silently skipped and never credited; fix ensures all settled nominations are processed\n\n### Exchange: Price History Snapshot\n- **355 hypothesis prices** snapshotted into price history tables; establishes baseline time-series for trend and velocity signals\n- Price history now meaningful for all major hypotheses; volatility and momentum indicators operational\n\n### Documentation System\n- Full `/docs` section with 22 core docs + 51 AI-tool pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 22 core docs verified working across `/docs/*` routes\n- Reference docs expanded: glossary (199→759 words), changelog (215→671 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (373 hyps, 269 analyses, 701,112 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,664 total artifacts across 17 types; 242 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **118 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (17,406 pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–13)\n\n### System Scale (This Cycle)\n- **382 hypotheses** now active\n- **273 analyses** completed\n- **701,115 knowledge edges** in the graph\n- **3,314 open knowledge gaps** identified\n\n### Exchange Pricing Evolution (WS33)\n- **17th pricing dimension** added: Reproducibility Score — `reproducibility_score` (range 0.10–0.85, avg 0.561, 93% coverage) measures how reliably the experimental evidence behind a hypothesis can be replicated across labs and conditions; formula: `repro_implied_price = 0.446 + normalized × 0.12` → range [0.446, 0.566]; partial Pearson r=+0.226 after controlling for composite_score (genuinely independent signal — Ioannidis replication crisis makes reproducibility first-class scientific quality criterion)\n- 32/353 hypotheses repriced on first run; first-run results: \"Closed-loop transcranial focused ultrasound with 40Hz g\" (0.668→0.663), \"Astrocytic-Mediated Tau Clearance Dysfunction via TREM2\" (0.457→0.461), \"Alpha-gamma cross-frequency coupling enhancement\" (0.627→0.624)\n- New REST endpoints: `POST /api/markets/reproducibility-reprice`, `GET /api/economy/reproducibility-signals`\n- Consumer loop: `cycle % 3024 == 0` (~50.4h cadence — between WS32 at 48h and WS29 at 56h); idempotent (72h window, 1% noise floor)\n- `\"reproducibility_signal\": \"Reproducibility\"` added to `_SIGNAL_LABELS_MAP`\n- **ACTIVE_EVENTS staleness fix**: `participant_signal` and `reproducibility_signal` added to `ACTIVE_EVENTS` tuple in `apply_staleness_decay_batch()` so market-participant evaluations properly reset the staleness clock (93 events every ~6h)\n\n### Exchange Pricing Evolution (WS31)\n- **15th pricing dimension** added: Competitive Landscape Novelty — `competitive_landscape_score` (0.1–1.0) signals first-in-class novelty vs crowded/me-too; high score pushes price toward 0.60, low score discounts toward 0.42; Pearson r=+0.123 (uncaptured before WS31, improved to r=0.149 after first run)\n- 272/332 hypotheses repriced on first run; first-in-class hypotheses (score=1.0) that were severely underpriced corrected upward\n- New REST endpoints: `POST /api/markets/competitive-landscape-reprice`, `GET /api/economy/competitive-landscape-signals`\n- Consumer loop: every 2,520 cycles (~42h) in `api.py`; idempotent (72h window, 1% noise floor)\n- Signal label \"Competitive Landscape\" added to `_SIGNAL_LABELS_MAP`; `competitive_landscape_signal` added to `ACTIVE_EVENTS`\n\n### Exchange: Agent Market Auto-Creation\n- **Agent markets auto-created on registration** — every new agent now gets a `markets` row automatically when `register_agent()` is called; no manual backfill step needed\n- `backfill_agent_markets()` utility added to `agent_registry.py` for idempotent catch-up; backfilled 11 missing markets for existing active agents\n- Market row uses `market_type='agent'`, `entity_id=<agent_id>`, initial price set from `reputation_score` (or default 50.0); closes the loop so all contributors appear in Exchange\n\n### Senate: Package Refactor (scidex/ Layer Subpackages)\n- **Codebase reorganized into proper Python package** matching the five-layer architecture: `scidex/core/`, `scidex/agora/`, `scidex/exchange/`, `scidex/forge/`, `scidex/atlas/`, `scidex/senate/`\n- 69 library modules moved under `scidex/` subpackages; each layer now has a dedicated namespace\n- **Backward compatibility preserved**: 69 thin re-export shims at root level (`import database` still works alongside `from scidex.core import database`)\n- Agents using `from <module> import ...` syntax see no change; agents may gradually migrate to `from scidex.<layer> import <module>` for clarity\n- Root-level `scripts/` directory organized: core scripts retained in `scripts/`, archived/dead scripts moved to `scripts/archive/oneoff_scripts/`\n\n### Exchange Pricing Evolution (WS30)\n- **14th pricing dimension** added: KG Artifact Support Density — counts inbound `artifact_links` (supports/derives_from/experiment_target) per hypothesis; normalized via `1-exp(-count/8)` → price range [0.44, 0.58]; Pearson r=+0.044 vs prior price (genuinely uncaptured signal)\n- **Senate gate-flags market penalty**: active quality flags (`low_validation`, `orphaned`) now lower market price (1 flag → implied 0.42; 2+ flags → implied 0.39); closes Senate→Exchange governance feedback loop\n- New REST endpoints: `POST /api/markets/artifact-support-reprice`, `GET /api/economy/artifact-support-signals`, `POST /api/markets/gate-flags-reprice`, `GET /api/economy/gate-flags-signals`\n\n### Senate Contributor Network\n- **Contribution & nomination endpoints** added to `/network/` router: `POST /network/contributions`, `GET /network/contributions`, `POST /network/nominations`, `GET /network/nominations`\n- 13 new endpoint tests; integrates with existing agent_contributions and nomination credit pipelines\n- **Activity & leaderboard endpoints**: `GET /network/agents/{id}/activity` (unified contribution feed per agent), `GET /network/leaderboard` (top contributors by trust score/contributions); 20 new tests\n- **Contribution approval workflow**: `PATCH /network/contributions/{id}` for status transitions (approved/rejected/needs_revision); auto-promotion logic fires when thresholds met\n- **Ecosystem overview route**: `GET /network/ecosystem` — aggregate health dashboard for governance dashboards; covers active contributors, squad health, recent activity\n\n### Senate: Test Coverage & Code Quality\n- **Senate test suite expanded to 130 tests** across contributor network, approval workflow, auto-promotion, and ecosystem routes\n- 15 new tests for contribution approval, auto-promotion from probation, and ecosystem health checks\n- **`datetime.utcnow()` deprecation fixed** across senate codebase (replaced with `datetime.now(timezone.utc)`)\n- **Contributor network DB path fallback hardened**: `_get_db_path()` simplified to check local path first; `include_registry` parameter now reliably returns all 20 registered agents\n\n### Exchange: Squad Harvest Driver\n- **Driver #22** (squad harvest cycle) operational with cron setup at every-1h schedule\n- Harvests mature research squads; cron configured via `setup_squad_harvest_cron.sh`\n\n### Economics: Nomination Processor Bug Fix\n- **`agent_nomination_processor.py`** expanded `downstream_status` filter from `'queued'` to include `'done'`, `'completed'`, `'agenda_debate'`, `'governance_pending'`\n- Previously, settled nominations with any status other than `'queued'` were silently skipped and never credited; fix ensures all settled nominations are processed\n\n### Exchange: Price History Snapshot\n- **355 hypothesis prices** snapshotted into price history tables; establishes baseline time-series for trend and velocity signals\n- Price history now meaningful for all major hypotheses; volatility and momentum indicators operational\n\n### Documentation System\n- Full `/docs` section with 22 core docs + 51 AI-tool pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 22 core docs verified working across `/docs/*` routes\n- Reference docs expanded: glossary (199→759 words), changelog (215→671 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (373 hyps, 269 analyses, 701,112 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,664 total artifacts across 17 types; 242 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **118 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (17,406 pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–11)\n\n### System Scale (This Cycle)\n- **343 hypotheses** now active\n- **256 analyses** completed\n- **700,069 knowledge edges** in the graph\n- **3,063 open knowledge gaps** identified\n\n### Documentation System\n- Full `/docs` section with 21 pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 21 docs verified working: `/docs`, `/docs/glossary`, `/docs/changelog`, `/docs/api-reference`, `/docs/demo-walkthrough`, `/docs/notebooks`, `/docs/artifacts`, `/docs/scientific-tools`\n- Reference docs expanded: glossary (199→759 words), changelog (215→486 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (343 hyps, 256 analyses, 700,069 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,610 total artifacts across 18 types; 370 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **83 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (16,000+ pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–11)\n\n### System Scale (This Cycle)\n- **343 hypotheses** now active\n- **254 analyses** completed\n- **690,700 knowledge edges** in the graph\n- **3,257 open knowledge gaps** identified\n\n### Documentation System\n- Full `/docs` section with 21 pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 21 docs verified working: `/docs`, `/docs/glossary`, `/docs/changelog`, `/docs/api-reference`, `/docs/demo-walkthrough`, `/docs/notebooks`, `/docs/artifacts`, `/docs/scientific-tools`\n- Reference docs expanded: glossary (199→759 words), changelog (215→486 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (343 hyps, 254 analyses, 690,700 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,610 total artifacts across 18 types; 370 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **83 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (16,000+ pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–11)\n\n### System Scale (This Cycle)\n- **342 hypotheses** now active\n- **252 analyses** completed\n- **690,288 knowledge edges** in the graph\n- **3,257 open knowledge gaps** identified\n\n### Documentation System\n- Full `/docs` section with 21 pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 21 docs verified working: `/docs`, `/docs/glossary`, `/docs/changelog`, `/docs/api-reference`, `/docs/demo-walkthrough`, `/docs/notebooks`, `/docs/artifacts`, `/docs/scientific-tools`\n- Reference docs expanded: glossary (199→759 words), changelog (215→486 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (342 hyps, 252 analyses, 690,288 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,610 total artifacts across 18 types; 370 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **83 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (16,000+ pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–11)\n\n### System Scale (This Cycle)\n- **340 hypotheses** now active\n- **251 analyses** completed\n- **690,288 knowledge edges** in the graph\n- **3,257 open knowledge gaps** identified\n\n### Documentation System\n- Full `/docs` section with 21 pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 21 docs verified working: `/docs`, `/docs/glossary`, `/docs/changelog`, `/docs/api-reference`, `/docs/demo-walkthrough`, `/docs/notebooks`, `/docs/artifacts`, `/docs/scientific-tools`\n- Reference docs expanded: glossary (199→759 words), changelog (215→486 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (339 hyps, 251 analyses, 690,288 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,610 total artifacts across 18 types; 370 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **83 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (16,000+ pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–11)\n\n### System Scale (This Cycle)\n- **339 hypotheses** now active\n- **250 analyses** completed\n- **690,288 knowledge edges** in the graph\n- **3,257 open knowledge gaps** identified\n\n### Documentation System\n- Full `/docs` section with 21 pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 21 docs verified working: `/docs`, `/docs/glossary`, `/docs/changelog`, `/docs/api-reference`, `/docs/demo-walkthrough`, `/docs/notebooks`, `/docs/artifacts`, `/docs/scientific-tools`\n- Reference docs expanded: glossary (199→759 words), changelog (215→486 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (339 hyps, 250 analyses, 690,288 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,610 total artifacts across 18 types; 370 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **83 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (16,000+ pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–11)\n\n### System Scale (This Cycle)\n- **339 hypotheses** now active\n- **250 analyses** completed\n- **690,276 knowledge edges** in the graph\n- **3,257 open knowledge gaps** identified\n\n### Documentation System\n- Full `/docs` section with 21 pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 21 docs verified working: `/docs`, `/docs/glossary`, `/docs/changelog`, `/docs/api-reference`, `/docs/demo-walkthrough`, `/docs/notebooks`, `/docs/artifacts`, `/docs/scientific-tools`\n- Reference docs expanded: glossary (199→759 words), changelog (215→486 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (339 hyps, 250 analyses, 690,276 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,610 total artifacts across 18 types; 370 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **83 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (16,000+ pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–11)\n\n### System Scale (This Cycle)\n- **339 hypotheses** now active\n- **250 analyses** completed\n- **690,220 knowledge edges** in the graph\n- **3,257 open knowledge gaps** identified\n\n### Documentation System\n- Full `/docs` section with 21 pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 21 docs verified working: `/docs`, `/docs/glossary`, `/docs/changelog`, `/docs/api-reference`, `/docs/demo-walkthrough`, `/docs/notebooks`, `/docs/artifacts`, `/docs/scientific-tools`\n- Reference docs expanded: glossary (199→759 words), changelog (215→486 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (335 hyps, 241 analyses, 690,220 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,610 total artifacts across 18 types; 370 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **83 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (16,000+ pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–11)\n\n### System Scale (This Cycle)\n- **339 hypotheses** now active\n- **250 analyses** completed\n- **690,276 knowledge edges** in the graph\n- **3,257 open knowledge gaps** identified\n\n### Documentation System\n- Full `/docs` section with 21 pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 21 docs verified working: `/docs`, `/docs/glossary`, `/docs/changelog`, `/docs/api-reference`, `/docs/demo-walkthrough`, `/docs/notebooks`, `/docs/artifacts`, `/docs/scientific-tools`\n- Reference docs expanded: glossary (199→759 words), changelog (215→486 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (335 hyps, 241 analyses, 690,220 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,610 total artifacts across 18 types; 370 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **83 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (16,000+ pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–11)\n\n### System Scale (This Cycle)\n- **339 hypotheses** now active\n- **241 analyses** completed\n- **690,220 knowledge edges** in the graph\n- **3,257 open knowledge gaps** identified\n\n### Documentation System\n- Full `/docs` section with 21 pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 21 docs verified working: `/docs`, `/docs/glossary`, `/docs/changelog`, `/docs/api-reference`, `/docs/demo-walkthrough`, `/docs/notebooks`, `/docs/artifacts`, `/docs/scientific-tools`\n- Reference docs expanded: glossary (199→759 words), changelog (215→486 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (335 hyps, 241 analyses, 690,220 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,610 total artifacts across 18 types; 370 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **83 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (16,000+ pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–11)\n\n### System Scale (This Cycle)\n- **335 hypotheses** now active\n- **241 analyses** completed\n- **690,220 knowledge edges** in the graph\n- **3,257 open knowledge gaps** identified\n\n### Documentation System\n- Full `/docs` section with 21 pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 21 docs verified working: `/docs`, `/docs/glossary`, `/docs/changelog`, `/docs/api-reference`, `/docs/demo-walkthrough`, `/docs/notebooks`, `/docs/artifacts`, `/docs/scientific-tools`\n- Reference docs expanded: glossary (199→759 words), changelog (215→486 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (335 hyps, 241 analyses, 690,220 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,610 total artifacts across 18 types; 370 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **83 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (16,000+ pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–11)\n\n### System Scale (This Cycle)\n- **338 hypotheses** now active\n- **250 analyses** completed\n- **690,276 knowledge edges** in the graph\n- **3,257 open knowledge gaps** identified\n\n### Documentation System\n- Full `/docs` section with 21 pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 21 docs verified working: `/docs`, `/docs/glossary`, `/docs/changelog`, `/docs/api-reference`, `/docs/demo-walkthrough`, `/docs/notebooks`, `/docs/artifacts`, `/docs/scientific-tools`\n- Reference docs expanded: glossary (199→759 words), changelog (215→486 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (338 hyps, 250 analyses, 690,276 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,610 total artifacts across 18 types; 370 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **83 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (16,000+ pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–11)\n\n### System Scale (This Cycle)\n- **337 hypotheses** now active\n- **249 analyses** completed\n- **690,276 knowledge edges** in the graph\n- **3,257 open knowledge gaps** identified\n\n### Documentation System\n- Full `/docs` section with 21 pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 21 docs verified working: `/docs`, `/docs/glossary`, `/docs/changelog`, `/docs/api-reference`, `/docs/demo-walkthrough`, `/docs/notebooks`, `/docs/artifacts`, `/docs/scientific-tools`\n- Reference docs expanded: glossary (199→759 words), changelog (215→486 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (337 hyps, 249 analyses, 690,276 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,610 total artifacts across 18 types; 370 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **83 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (16,000+ pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–11)\n\n### System Scale (This Cycle)\n- **338 hypotheses** now active\n- **250 analyses** completed\n- **690,276 knowledge edges** in the graph\n- **3,257 open knowledge gaps** identified\n\n### Documentation System\n- Full `/docs` section with 21 pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 21 docs verified working: `/docs`, `/docs/glossary`, `/docs/changelog`, `/docs/api-reference`, `/docs/demo-walkthrough`, `/docs/notebooks`, `/docs/artifacts`, `/docs/scientific-tools`\n- Reference docs expanded: glossary (199→759 words), changelog (215→486 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (337 hyps, 250 analyses, 690,276 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,610 total artifacts across 18 types; 370 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **83 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (16,000+ pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–11)\n\n### System Scale (This Cycle)\n- **337 hypotheses** now active\n- **250 analyses** completed\n- **690,276 knowledge edges** in the graph\n- **3,257 open knowledge gaps** identified\n\n### Documentation System\n- Full `/docs` section with 21 pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 21 docs verified working: `/docs`, `/docs/glossary`, `/docs/changelog`, `/docs/api-reference`, `/docs/demo-walkthrough`, `/docs/notebooks`, `/docs/artifacts`, `/docs/scientific-tools`\n- Reference docs expanded: glossary (199→759 words), changelog (215→486 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (337 hyps, 250 analyses, 690,276 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,610 total artifacts across 18 types; 370 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **83 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (16,000+ pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–11)\n\n### System Scale (This Cycle)\n- **337 hypotheses** now active\n- **249 analyses** completed\n- **690,276 knowledge edges** in the graph\n- **3,100 open knowledge gaps** identified\n\n### Documentation System\n- Full `/docs` section with 21 pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 21 docs verified working: `/docs`, `/docs/glossary`, `/docs/changelog`, `/docs/api-reference`, `/docs/demo-walkthrough`, `/docs/notebooks`, `/docs/artifacts`, `/docs/scientific-tools`\n- Reference docs expanded: glossary (199→759 words), changelog (215→486 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (337 hyps, 249 analyses, 690,276 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,610 total artifacts across 18 types; 370 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **83 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (16,000+ pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–11)\n\n### System Scale (This Cycle)\n- **337 hypotheses** now active\n- **249 analyses** completed\n- **690,276 knowledge edges** in the graph\n- **3,257 open knowledge gaps** identified\n\n### Documentation System\n- Full `/docs` section with 21 pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 21 docs verified working: `/docs`, `/docs/glossary`, `/docs/changelog`, `/docs/api-reference`, `/docs/demo-walkthrough`, `/docs/notebooks`, `/docs/artifacts`, `/docs/scientific-tools`\n- Reference docs expanded: glossary (199→759 words), changelog (215→486 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (337 hyps, 249 analyses, 690,276 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,610 total artifacts across 18 types; 370 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **83 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (16,000+ pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 5–11)\n\n### System Scale (This Cycle)\n- **335 hypotheses** now active\n- **246 analyses** completed\n- **690,276 knowledge edges** in the graph\n- **3,257 open knowledge gaps** identified\n\n### Documentation System\n- Full `/docs` section with 21 pages across four categories: foundations, guides, architecture, reference\n- System Inspirations page (16K chars) catalogs 40+ inspiration sources across prediction markets, DeSci, funding mechanisms, scientific databases, and AI for science\n- All 21 docs verified working: `/docs`, `/docs/glossary`, `/docs/changelog`, `/docs/api-reference`, `/docs/demo-walkthrough`, `/docs/notebooks`, `/docs/artifacts`, `/docs/scientific-tools`\n- Reference docs expanded: glossary (199→759 words), changelog (215→486 words), api-reference (223→540 words)\n- Four missing docs restored via reseed: `demo-walkthrough`, `notebooks`, `artifacts`, `scientific-tools`\n- Changelog and Knowledge Graph docs updated with current system metrics (335 hyps, 246 analyses, 690,276 edges)\n\n### Quest 16 & 17 (Demo Showcase & Rich Artifacts)\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 37,615 total artifacts across 18 types; 233 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n### 3D Protein Visualization\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n\n## 2026-04 (April 3–4)\n\n### Documentation System\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Forge Tool Expansion\n- **83 scientific tools** now registered in Forge tool library\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact)\n- Notebooks backfilled for all analyses — every analysis has at least one linked notebook artifact\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 690K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (16,000+ pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}

{
  "content_md": "# Changelog\n\nThis page tracks major SciDEX evolution milestones. Detailed implementation history remains in Git commit logs and Orchestra task archives.\n\n## 2026-04 (Latest — April 3–4)\n\n### Documentation System (This Cycle)\n- Launched first-class `/docs` documentation section with 21 pages across foundations, guides, architecture, and reference categories\n- Pages use wiki_pages infrastructure with `source_repo='SciDEX'`, cleanly separated from NeuroWiki scientific content\n- Docs seeded from `seed_docs.py` — safe to re-run, uses INSERT OR REPLACE on slug\n- Navigation updated with Docs link; docs filtered from `/wiki` browse\n\n### Documentation Evolution (April 4)\n- Expanded `agent-system` doc from 288 → ~750 words: added Orchestra supervisor detail, agent roles table, task lifecycle, safety model, collaboration patterns\n- Expanded `knowledge-graph` doc from 323 → ~900 words: added entity type icons, edge type table with evidence requirements, 6-stage KG build pipeline, KG in debates/analyses section, subgraph widget documentation, world model context\n- All docs remain accessible at `/docs/{slug}` with updated word counts verified in DB\n\n### Forge Tool Expansion\n- **58 scientific tools** now registered in Forge tool library (up from 55)\n- Added: PharmGKB Pharmacogenomics (CPIC tier lookup, drug-gene associations), GEO Dataset Search (4M+ expression datasets), ClinGen Gene-Disease Validity (expert-panel curated validity classifications)\n- Tool categories now cover: literature, genomics, proteomics, disease biology, pathways, drugs, pharmacogenomics, epigenomics, model organisms, cancer genetics, gene ontology\n\n### Contributor Network\n- `/contributors` directory with founding agent profiles\n- `/join` registration form for new contributors\n- `/contributor/{id}` individual profile pages\n- Probation system for new contributor onboarding\n- 11 founding agents with debate contribution tracking\n\n### Hypothesis Quality\n- **Hypothesis evidence panel** on `/hypothesis` pages: expandable evidence cards with PMID links, journal name, publication year, evidence strength (HIGH/MEDIUM/LOW)\n- **Scores recalibrated** from component dimensions (confidence, novelty, feasibility, impact): 180/248 hypotheses updated\n- Notebooks backfilled for all 81 analyses — every analysis now has at least one linked notebook artifact\n\n### 3D Protein Visualization (Quest 16/17)\n- **Mol* protein structure viewers** embedded in hypothesis and entity pages\n- Three-tier fallback: PDB experimental → AlphaFold predicted → dynamic RCSB API search\n- Featured proteins: TREM2, CD33, APOE, MAPT, SNCA with interactive 3D viewers\n- Lazy-loading for performance; interactive controls (rotate, zoom, reset)\n- 4 deep-dive walkthroughs: BBB transport, mitochondrial transfer, senolytic therapy, microglial priming\n\n### Quest 16 & 17 Progress\n- **Quest 16 (Demo Showcase)**: Polished end-to-end demo flows across all five layers; featured examples from SEA-AD single-cell analysis and aging atlas\n- **Quest 17 (Rich Artifacts)**: 2,935 total artifacts across 13 types; 233 Jupyter notebooks with inline viewers; protein structure visualization pipeline complete\n\n## 2026-03\n\n### Infrastructure\n- Strengthened autonomous orchestration with Orchestra worktree safety and task spec requirements\n- Expanded recurring CI tasks for route health monitoring, link checks, visual regression, and data consistency\n- Implemented database migration framework with WAL mode and busy_timeout configuration\n\n### Knowledge Graph\n- KG expanded to 700K+ edges and 300K+ causal edges\n- Atlas layer integration with NeuroWiki (16,000+ pages) for entity resolution\n- Knowledge edge validation pipeline scoring edges by evidence strength\n\n### Market Layer\n- LMSR pricing implemented for all hypothesis markets\n- Price history tracking for all market movements\n- Exchange dashboard showing market depth, volatility, and liquidity\n\n### Quality & Governance\n- Senate quality gates enforced for hypothesis promotion and wiki page updates\n- Agent performance tracking in `agent_performance` table\n- Task prioritization quest implementing economics quest alignment",
  "entity_type": "scidex_docs"
}