{"quest":{"id":"q-epistemic-rigor","name":"Epistemic Rigor","description":"Evolve SciDEX toward rigorous scientific epistemology: 6-tier epistemic classification (T0 axiom → T5 contested) with asymmetric promotion/demotion (Gödel Machine principle), adversarial Falsifier persona as 5th debate round, proof-gated KG updates with contradiction checking against established knowledge, GFlowNet-inspired diverse hypothesis portfolio maintenance (anti-monoculture rules), pre-registration of predictions before analysis, replication status tracking (unreplicated/replicated/failed_replication), and calibration-adjusted believability weights. Every claim must be traceable to ground truth.","layer":"Cross-cutting","priority":95,"status":"active","created_at":"2026-04-03 23:02:52","updated_at":"2026-04-03 23:09:38"},"tasks":[{"id":"t-epistemic-promotion-ui","title":"Display epistemic tier promotion roadmap on hypothesis pages","description":"","status":"open","priority":30,"task_type":"one_shot","frequency":"","assigned_slot":"","started_at":null,"completed_at":null,"updated_at":"2026-04-25T07:47:19.383916+00:00","summary":"","completion_notes":"","last_error":"cli-reopen-manual: reopened — task was marked 'done' but has no task_runs row in (done/completed/success)","time_estimate_hours":0.0,"completion_count":0,"spec_path":"","provider":"claude","payload_json":"{}"},{"id":"de85e156-b7e3-4513-a586-4f3762a824f7","title":"[Atlas] Hypothesis predictions table — explicit falsifiability","description":"Add hypothesis_predictions table: each hypothesis gets explicit testable predictions with predicted outcomes, falsification criteria, and status tracking. Add falsifiable flag + predictions_count to hypotheses. Backfill existing hypotheses by extracting predictions from debate transcripts.\n\n\n## REOPENED TASK — CRITICAL CONTEXT\n\nThis task was previously marked 'done' but the audit could not verify\nthe work actually landed on main. The original work may have been:\n- Lost to an orphan branch / failed push\n- Only a spec-file edit (no code changes)\n- Already addressed by other agents in the meantime\n- Made obsolete by subsequent work\n\n**Before doing anything else:**\n\n1. **Re-evaluate the task in light of CURRENT main state.** Read the\n spec and the relevant files on origin/main NOW. The original task\n may have been written against a state of the code that no longer\n exists.\n\n2. **Verify the task still advances SciDEX's aims.** If the system\n has evolved past the need for this work (different architecture,\n different priorities), close the task with reason \"obsolete: \"\n instead of doing it.\n\n3. **Check if it's already done.** Run `git log --grep=''`\n and read the related commits. If real work landed, complete the\n task with `--no-sha-check --summary 'Already done in '`.\n\n4. **Make sure your changes don't regress recent functionality.** Many\n agents have been working on this codebase. Before committing, run\n `git log --since='24 hours ago' -- ` to see what\n changed in your area, and verify you don't undo any of it.\n\n5. **Stay scoped.** Only do what this specific task asks for. Do not\n refactor, do not \"fix\" unrelated issues, do not add features that\n weren't requested. Scope creep at this point is regression risk.\n\nIf you cannot do this task safely (because it would regress, conflict\nwith current direction, or the requirements no longer apply), escalate\nvia `orchestra escalate` with a clear explanation instead of committing.\n","status":"done","priority":95,"task_type":"one_shot","frequency":"","assigned_slot":"","started_at":null,"completed_at":"2026-04-15T23:42:22.095910+00:00","updated_at":"2026-04-15T23:42:22.095910+00:00","summary":"","completion_notes":"Auto-completed by supervisor after successful deploy to main","last_error":"","time_estimate_hours":0.0,"completion_count":0,"spec_path":"docs/planning/specs/de85e156_b7e_spec.md","provider":"any","payload_json":"{\"requirements\": {\"analysis\": 7, \"reasoning\": 6}, \"_stall_skip_providers\": [], \"_stall_requeued_by\": \"minimax\", \"_stall_requeued_at\": \"2026-04-14 00:16:29\", \"_stall_skip_at\": {}, \"_stall_skip_pruned_at\": \"2026-04-14T10:37:14.022390+00:00\", \"completion_shas\": [\"7abb57f3ea325a62100e7bf6bf3f13aae09c99fd\"], \"completion_shas_checked_at\": \"2026-04-15T23:42:22.077420+00:00\"}"},{"id":"5f27f904-d33b-4f57-9c05-110d93f20ad9","title":"[Forge] Experiment results and validation pipeline","description":"Add experiment_results table linking experiment outcomes to predictions. When an experiment completes, compare actual results to predicted outcomes and auto-update hypothesis scores. Build the prediction-vs-reality comparison engine. Connect to existing experiments table.\n\n\n## REOPENED TASK — CRITICAL CONTEXT\n\nThis task was previously marked 'done' but the audit could not verify\nthe work actually landed on main. The original work may have been:\n- Lost to an orphan branch / failed push\n- Only a spec-file edit (no code changes)\n- Already addressed by other agents in the meantime\n- Made obsolete by subsequent work\n\n**Before doing anything else:**\n\n1. **Re-evaluate the task in light of CURRENT main state.** Read the\n spec and the relevant files on origin/main NOW. The original task\n may have been written against a state of the code that no longer\n exists.\n\n2. **Verify the task still advances SciDEX's aims.** If the system\n has evolved past the need for this work (different architecture,\n different priorities), close the task with reason \"obsolete: \"\n instead of doing it.\n\n3. **Check if it's already done.** Run `git log --grep=''`\n and read the related commits. If real work landed, complete the\n task with `--no-sha-check --summary 'Already done in '`.\n\n4. **Make sure your changes don't regress recent functionality.** Many\n agents have been working on this codebase. Before committing, run\n `git log --since='24 hours ago' -- ` to see what\n changed in your area, and verify you don't undo any of it.\n\n5. **Stay scoped.** Only do what this specific task asks for. Do not\n refactor, do not \"fix\" unrelated issues, do not add features that\n weren't requested. Scope creep at this point is regression risk.\n\nIf you cannot do this task safely (because it would regress, conflict\nwith current direction, or the requirements no longer apply), escalate\nvia `orchestra escalate` with a clear explanation instead of committing.\n","status":"done","priority":93,"task_type":"one_shot","frequency":"","assigned_slot":"","started_at":null,"completed_at":"2026-04-14T04:09:57.790000+00:00","updated_at":"2026-04-14T04:09:57.790000+00:00","summary":"","completion_notes":"Auto-completed by supervisor after successful deploy to main","last_error":"","time_estimate_hours":0.0,"completion_count":0,"spec_path":"docs/planning/specs/5f27f904_d33_spec.md","provider":"any","payload_json":"{\"requirements\": {\"coding\": 8, \"safety\": 9}, \"completion_shas\": [\"ca83f0ce0d5abf9ba16e8a7e2c86aa2741fe8644\", \"cfa06e56a685f6f762b27697930d92f3502c39b8\"], \"completion_shas_checked_at\": \"2026-04-14T04:09:57.770302+00:00\", \"completion_shas_missing\": [\"641517efac9d1628df1b58395d04526bc467ce66\", \"3d4260ea9b9702ecf33ca0ac743cac06deb9eb4a\"]}"},{"id":"dca29466-2f7e-4a1e-848b-1049acc67ffc","title":"[Agora] epi-01-PRED: Add hypothesis_predictions table for falsifiable predictions","description":"Create hypothesis_predictions table linking hypotheses to explicit testable predictions. Each prediction should have: the claim, what would falsify it, expected measurement, confidence level. This is the foundational step for epistemic rigor — transforms hypotheses from scored claims to scientifically testable statements. Spec: quest_epistemic_rigor.md\n\n\n## REOPENED TASK — CRITICAL CONTEXT\n\nThis task was previously marked 'done' but the audit could not verify\nthe work actually landed on main. The original work may have been:\n- Lost to an orphan branch / failed push\n- Only a spec-file edit (no code changes)\n- Already addressed by other agents in the meantime\n- Made obsolete by subsequent work\n\n**Before doing anything else:**\n\n1. **Re-evaluate the task in light of CURRENT main state.** Read the\n spec and the relevant files on origin/main NOW. The original task\n may have been written against a state of the code that no longer\n exists.\n\n2. **Verify the task still advances SciDEX's aims.** If the system\n has evolved past the need for this work (different architecture,\n different priorities), close the task with reason \"obsolete: \"\n instead of doing it.\n\n3. **Check if it's already done.** Run `git log --grep=''`\n and read the related commits. If real work landed, complete the\n task with `--no-sha-check --summary 'Already done in '`.\n\n4. **Make sure your changes don't regress recent functionality.** Many\n agents have been working on this codebase. Before committing, run\n `git log --since='24 hours ago' -- ` to see what\n changed in your area, and verify you don't undo any of it.\n\n5. **Stay scoped.** Only do what this specific task asks for. Do not\n refactor, do not \"fix\" unrelated issues, do not add features that\n weren't requested. Scope creep at this point is regression risk.\n\nIf you cannot do this task safely (because it would regress, conflict\nwith current direction, or the requirements no longer apply), escalate\nvia `orchestra escalate` with a clear explanation instead of committing.\n","status":"done","priority":92,"task_type":"one_shot","frequency":"","assigned_slot":"","started_at":null,"completed_at":"2026-04-16T09:55:07.071394+00:00","updated_at":"2026-04-16T09:55:07.071394+00:00","summary":"","completion_notes":"Audit misreported ORPHAN_BRANCH. Commit 835ef91f7 ([Agora] epi-01-PRED: Add hypothesis_predictions columns, API endpoints, and detail page section [task:dca29466-2f7e-4a1e-848b-1049acc67ffc]) IS on origin/main - verified via git merge-base --is-ancestor 835ef91f7 82b003414 (origin/main HEAD). Migration 056 applied, table has 988 rows, API endpoints functional.","last_error":"","time_estimate_hours":0.0,"completion_count":0,"spec_path":"docs/planning/specs/quest_epistemic_rigor.md","provider":"any","payload_json":"{\"requirements\": {\"analysis\": 7, \"reasoning\": 6}, \"completion_shas\": [\"835ef91f7c285957828dc04ab4c418be72b6cea3\"], \"completion_shas_checked_at\": \"2026-04-16T09:55:07.049627+00:00\"}"},{"id":"b5298ea7-69ec-47bd-9115-f2968a374f6d","title":"[Atlas] Evidence chain provenance — trace every claim to ground truth","description":"Add evidence_chains table: tracks full provenance of every evidence claim (who evaluated it, when, using what methodology, from what source). Normalize evidence_for/evidence_against from JSON into structured evidence_entries table with source_type (paper/experiment/debate/manual), source_id, methodology, evaluator_agent_id. Every claim traceable to ground data.\n\n\n## REOPENED TASK — CRITICAL CONTEXT\n\nThis task was previously marked 'done' but the audit could not verify\nthe work actually landed on main. The original work may have been:\n- Lost to an orphan branch / failed push\n- Only a spec-file edit (no code changes)\n- Already addressed by other agents in the meantime\n- Made obsolete by subsequent work\n\n**Before doing anything else:**\n\n1. **Re-evaluate the task in light of CURRENT main state.** Read the\n spec and the relevant files on origin/main NOW. The original task\n may have been written against a state of the code that no longer\n exists.\n\n2. **Verify the task still advances SciDEX's aims.** If the system\n has evolved past the need for this work (different architecture,\n different priorities), close the task with reason \"obsolete: \"\n instead of doing it.\n\n3. **Check if it's already done.** Run `git log --grep=''`\n and read the related commits. If real work landed, complete the\n task with `--no-sha-check --summary 'Already done in '`.\n\n4. **Make sure your changes don't regress recent functionality.** Many\n agents have been working on this codebase. Before committing, run\n `git log --since='24 hours ago' -- ` to see what\n changed in your area, and verify you don't undo any of it.\n\n5. **Stay scoped.** Only do what this specific task asks for. Do not\n refactor, do not \"fix\" unrelated issues, do not add features that\n weren't requested. Scope creep at this point is regression risk.\n\nIf you cannot do this task safely (because it would regress, conflict\nwith current direction, or the requirements no longer apply), escalate\nvia `orchestra escalate` with a clear explanation instead of committing.\n","status":"done","priority":92,"task_type":"one_shot","frequency":"","assigned_slot":"","started_at":null,"completed_at":"2026-04-16T07:50:34.493467+00:00","updated_at":"2026-04-16T07:50:34.493467+00:00","summary":"","completion_notes":"Auto-completed by supervisor after successful deploy to main","last_error":"","time_estimate_hours":0.0,"completion_count":0,"spec_path":"docs/planning/specs/b5298ea7_69e_spec.md","provider":"any","payload_json":"{\"completion_shas\": [\"891030ead9bce5da2a1be0a371fe3fb0381acb19\", \"a62e2c9186ce5e2876cea8d7bfcdf316e9f95f84\", \"fbc87b49cf5617a5df6808f6d9cecd51f84eb012\", \"d68545ac31d0391e86ff431b8a30aa4a8163a982\"], \"completion_shas_checked_at\": \"2026-04-16T07:50:34.472350+00:00\"}"},{"id":"edd66643-1332-4e67-b11e-0da9fadebaab","title":"[Atlas] Trust scores on knowledge graph edges","description":"Add trust_score (0-1) with confidence_interval to knowledge_edges. Compute trust from: evidence chain depth, source paper quality (impact factor, citations, recency), number of independent confirmations, contradiction count. Add trust propagation — downstream edges inherit reduced trust from upstream. Visualize trust on /graph.\n\n\n## REOPENED TASK — CRITICAL CONTEXT\n\nThis task was previously marked 'done' but the audit could not verify\nthe work actually landed on main. The original work may have been:\n- Lost to an orphan branch / failed push\n- Only a spec-file edit (no code changes)\n- Already addressed by other agents in the meantime\n- Made obsolete by subsequent work\n\n**Before doing anything else:**\n\n1. **Re-evaluate the task in light of CURRENT main state.** Read the\n spec and the relevant files on origin/main NOW. The original task\n may have been written against a state of the code that no longer\n exists.\n\n2. **Verify the task still advances SciDEX's aims.** If the system\n has evolved past the need for this work (different architecture,\n different priorities), close the task with reason \"obsolete: \"\n instead of doing it.\n\n3. **Check if it's already done.** Run `git log --grep=''`\n and read the related commits. If real work landed, complete the\n task with `--no-sha-check --summary 'Already done in '`.\n\n4. **Make sure your changes don't regress recent functionality.** Many\n agents have been working on this codebase. Before committing, run\n `git log --since='24 hours ago' -- ` to see what\n changed in your area, and verify you don't undo any of it.\n\n5. **Stay scoped.** Only do what this specific task asks for. Do not\n refactor, do not \"fix\" unrelated issues, do not add features that\n weren't requested. Scope creep at this point is regression risk.\n\nIf you cannot do this task safely (because it would regress, conflict\nwith current direction, or the requirements no longer apply), escalate\nvia `orchestra escalate` with a clear explanation instead of committing.\n","status":"done","priority":91,"task_type":"one_shot","frequency":"","assigned_slot":"","started_at":null,"completed_at":"2026-04-16T11:25:26.640847+00:00","updated_at":"2026-04-16T11:25:26.640847+00:00","summary":"","completion_notes":"Auto-completed by supervisor after successful deploy to main","last_error":"","time_estimate_hours":0.0,"completion_count":0,"spec_path":"docs/planning/specs/edd66643_133_spec.md","provider":"any","payload_json":"{\"requirements\": {\"coding\": 5, \"analysis\": 5}}"},{"id":"50b2dc6c-7208-4fd6-aae4-b997507e8515","title":"[Atlas] Hypothesis and experiment dependency graph","description":"Add hypothesis_dependencies table (hypothesis A depends on hypothesis B being true). Add experiment_prerequisites (experiment A must complete before B). Add evidence_supports (evidence X supports/contradicts claim Y). Build DAG traversal to compute critical paths — which experiments would most advance knowledge. Expose dependency graph via API and visualize.\n\n\n## REOPENED TASK — CRITICAL CONTEXT\n\nThis task was previously marked 'done' but the audit could not verify\nthe work actually landed on main. The original work may have been:\n- Lost to an orphan branch / failed push\n- Only a spec-file edit (no code changes)\n- Already addressed by other agents in the meantime\n- Made obsolete by subsequent work\n\n**Before doing anything else:**\n\n1. **Re-evaluate the task in light of CURRENT main state.** Read the\n spec and the relevant files on origin/main NOW. The original task\n may have been written against a state of the code that no longer\n exists.\n\n2. **Verify the task still advances SciDEX's aims.** If the system\n has evolved past the need for this work (different architecture,\n different priorities), close the task with reason \"obsolete: \"\n instead of doing it.\n\n3. **Check if it's already done.** Run `git log --grep=''`\n and read the related commits. If real work landed, complete the\n task with `--no-sha-check --summary 'Already done in '`.\n\n4. **Make sure your changes don't regress recent functionality.** Many\n agents have been working on this codebase. Before committing, run\n `git log --since='24 hours ago' -- ` to see what\n changed in your area, and verify you don't undo any of it.\n\n5. **Stay scoped.** Only do what this specific task asks for. Do not\n refactor, do not \"fix\" unrelated issues, do not add features that\n weren't requested. Scope creep at this point is regression risk.\n\nIf you cannot do this task safely (because it would regress, conflict\nwith current direction, or the requirements no longer apply), escalate\nvia `orchestra escalate` with a clear explanation instead of committing.\n","status":"done","priority":90,"task_type":"one_shot","frequency":"","assigned_slot":"","started_at":null,"completed_at":"2026-04-18T15:44:03.090168+00:00","updated_at":"2026-04-18T15:44:03.090168+00:00","summary":"","completion_notes":"Auto-completed by supervisor after successful deploy to main","last_error":"","time_estimate_hours":0.0,"completion_count":0,"spec_path":"docs/planning/specs/50b2dc6c_720_spec.md","provider":"any","payload_json":"{\"_reset_note\": \"This task was reset after a database incident on 2026-04-17.\\n\\n**Context:** SciDEX migrated from SQLite to PostgreSQL after recurring DB\\ncorruption. Some work done during Apr 16-17 may have been lost.\\n\\n**Before starting work:**\\n1. Check if the task's goal is ALREADY satisfied (run the relevant checks)\\n2. Check `git log --all --grep=task:YOUR_TASK_ID` for prior commits\\n3. If complete, verify and mark done. If partial, continue. If not done, proceed.\\n\\n**DB change:** SciDEX now uses PostgreSQL. `get_db()` auto-detects via\\nSCIDEX_DB_BACKEND=postgres env var.\", \"_reset_at\": \"2026-04-18T06:29:22.046013+00:00\", \"_reset_from_status\": \"done\"}"},{"id":"31ab7f59-adf4-4092-828f-ddf11b609d8f","title":"[Senate] Evidence versioning and audit trail","description":"Add evidence_versions table tracking every change to evidence, scores, and claims with timestamps, actor, reason, and diff. Add confidence_justifications table — every score change must have a structured reason (new evidence, debate outcome, experiment result, recalibration). Build audit trail viewer on hypothesis detail pages showing full history.\n\n\n## REOPENED TASK — CRITICAL CONTEXT\n\nThis task was previously marked 'done' but the audit could not verify\nthe work actually landed on main. The original work may have been:\n- Lost to an orphan branch / failed push\n- Only a spec-file edit (no code changes)\n- Already addressed by other agents in the meantime\n- Made obsolete by subsequent work\n\n**Before doing anything else:**\n\n1. **Re-evaluate the task in light of CURRENT main state.** Read the\n spec and the relevant files on origin/main NOW. The original task\n may have been written against a state of the code that no longer\n exists.\n\n2. **Verify the task still advances SciDEX's aims.** If the system\n has evolved past the need for this work (different architecture,\n different priorities), close the task with reason \"obsolete: \"\n instead of doing it.\n\n3. **Check if it's already done.** Run `git log --grep=''`\n and read the related commits. If real work landed, complete the\n task with `--no-sha-check --summary 'Already done in '`.\n\n4. **Make sure your changes don't regress recent functionality.** Many\n agents have been working on this codebase. Before committing, run\n `git log --since='24 hours ago' -- ` to see what\n changed in your area, and verify you don't undo any of it.\n\n5. **Stay scoped.** Only do what this specific task asks for. Do not\n refactor, do not \"fix\" unrelated issues, do not add features that\n weren't requested. Scope creep at this point is regression risk.\n\nIf you cannot do this task safely (because it would regress, conflict\nwith current direction, or the requirements no longer apply), escalate\nvia `orchestra escalate` with a clear explanation instead of committing.\n","status":"done","priority":88,"task_type":"one_shot","frequency":"","assigned_slot":"","started_at":null,"completed_at":"2026-04-19T00:56:15.329089+00:00","updated_at":"2026-04-19T00:56:15.329089+00:00","summary":"","completion_notes":"","last_error":"","time_estimate_hours":0.0,"completion_count":0,"spec_path":"docs/planning/specs/31ab7f59_adf_spec.md","provider":"any","payload_json":"{\"requirements\": {\"coding\": 8, \"safety\": 9}, \"_reset_note\": \"This task was reset after a database incident on 2026-04-17.\\n\\n**Context:** SciDEX migrated from SQLite to PostgreSQL after recurring DB\\ncorruption. Some work done during Apr 16-17 may have been lost.\\n\\n**Before starting work:**\\n1. Check if the task's goal is ALREADY satisfied (run the relevant checks)\\n2. Check `git log --all --grep=task:YOUR_TASK_ID` for prior commits\\n3. If complete, verify and mark done. If partial, continue. If not done, proceed.\\n\\n**DB change:** SciDEX now uses PostgreSQL. `get_db()` auto-detects via\\nSCIDEX_DB_BACKEND=postgres env var.\", \"_reset_at\": \"2026-04-18T06:29:22.046013+00:00\", \"_reset_from_status\": \"done\"}"},{"id":"08c73de3-03cc-481d-abb1-3bfa4a814ec1","title":"[Exchange] Combinable knowledge units — composable evidence blocks","description":"Design and implement knowledge_units — atomic, addressable, combinable evidence blocks. Each unit has: content, source provenance, trust score, version, links to hypotheses/experiments/papers. Units can be combined (A+B supports C), split, and reused across hypotheses. Build the foundation for reproducible, composable scientific reasoning.\n\n\n## REOPENED TASK — CRITICAL CONTEXT\n\nThis task was previously marked 'done' but the audit could not verify\nthe work actually landed on main. The original work may have been:\n- Lost to an orphan branch / failed push\n- Only a spec-file edit (no code changes)\n- Already addressed by other agents in the meantime\n- Made obsolete by subsequent work\n\n**Before doing anything else:**\n\n1. **Re-evaluate the task in light of CURRENT main state.** Read the\n spec and the relevant files on origin/main NOW. The original task\n may have been written against a state of the code that no longer\n exists.\n\n2. **Verify the task still advances SciDEX's aims.** If the system\n has evolved past the need for this work (different architecture,\n different priorities), close the task with reason \"obsolete: \"\n instead of doing it.\n\n3. **Check if it's already done.** Run `git log --grep=''`\n and read the related commits. If real work landed, complete the\n task with `--no-sha-check --summary 'Already done in '`.\n\n4. **Make sure your changes don't regress recent functionality.** Many\n agents have been working on this codebase. Before committing, run\n `git log --since='24 hours ago' -- ` to see what\n changed in your area, and verify you don't undo any of it.\n\n5. **Stay scoped.** Only do what this specific task asks for. Do not\n refactor, do not \"fix\" unrelated issues, do not add features that\n weren't requested. Scope creep at this point is regression risk.\n\nIf you cannot do this task safely (because it would regress, conflict\nwith current direction, or the requirements no longer apply), escalate\nvia `orchestra escalate` with a clear explanation instead of committing.\n","status":"done","priority":85,"task_type":"one_shot","frequency":"","assigned_slot":"","started_at":null,"completed_at":"2026-04-19T04:27:18.042106+00:00","updated_at":"2026-04-19T04:27:18.042106+00:00","summary":"","completion_notes":"Auto-completed by supervisor after successful deploy to main","last_error":"","time_estimate_hours":0.0,"completion_count":0,"spec_path":"docs/planning/specs/08c73de3_03c_spec.md","provider":"any","payload_json":"{\"_reset_note\": \"This task was reset after a database incident on 2026-04-17.\\n\\n**Context:** SciDEX migrated from SQLite to PostgreSQL after recurring DB\\ncorruption. Some work done during Apr 16-17 may have been lost.\\n\\n**Before starting work:**\\n1. Check if the task's goal is ALREADY satisfied (run the relevant checks)\\n2. Check `git log --all --grep=task:YOUR_TASK_ID` for prior commits\\n3. If complete, verify and mark done. If partial, continue. If not done, proceed.\\n\\n**DB change:** SciDEX now uses PostgreSQL. `get_db()` auto-detects via\\nSCIDEX_DB_BACKEND=postgres env var.\", \"_reset_at\": \"2026-04-18T06:29:22.046013+00:00\", \"_reset_from_status\": \"done\"}"},{"id":"4bb367b9-9d69-4807-a215-01f4c3323007","title":"[Senate] Epistemic health dashboard and continuous improvement","description":"Build /epistemic dashboard showing: falsifiability coverage (% hypotheses with predictions), evidence provenance coverage (% claims traced to source), trust score distribution across KG, dependency graph health, evidence freshness. Add Senate proposals that auto-flag hypotheses missing predictions, stale evidence, or broken provenance chains. This is the self-improvement loop for epistemic rigor.\n\n\n## REOPENED TASK — CRITICAL CONTEXT\n\nThis task was previously marked 'done' but the audit could not verify\nthe work actually landed on main. The original work may have been:\n- Lost to an orphan branch / failed push\n- Only a spec-file edit (no code changes)\n- Already addressed by other agents in the meantime\n- Made obsolete by subsequent work\n\n**Before doing anything else:**\n\n1. **Re-evaluate the task in light of CURRENT main state.** Read the\n spec and the relevant files on origin/main NOW. The original task\n may have been written against a state of the code that no longer\n exists.\n\n2. **Verify the task still advances SciDEX's aims.** If the system\n has evolved past the need for this work (different architecture,\n different priorities), close the task with reason \"obsolete: \"\n instead of doing it.\n\n3. **Check if it's already done.** Run `git log --grep=''`\n and read the related commits. If real work landed, complete the\n task with `--no-sha-check --summary 'Already done in '`.\n\n4. **Make sure your changes don't regress recent functionality.** Many\n agents have been working on this codebase. Before committing, run\n `git log --since='24 hours ago' -- ` to see what\n changed in your area, and verify you don't undo any of it.\n\n5. **Stay scoped.** Only do what this specific task asks for. Do not\n refactor, do not \"fix\" unrelated issues, do not add features that\n weren't requested. Scope creep at this point is regression risk.\n\nIf you cannot do this task safely (because it would regress, conflict\nwith current direction, or the requirements no longer apply), escalate\nvia `orchestra escalate` with a clear explanation instead of committing.\n","status":"done","priority":83,"task_type":"one_shot","frequency":"","assigned_slot":"","started_at":null,"completed_at":"2026-04-19T05:06:00.146624+00:00","updated_at":"2026-04-19T05:06:00.146624+00:00","summary":"","completion_notes":"Auto-completed by supervisor after successful deploy to main","last_error":"","time_estimate_hours":0.0,"completion_count":0,"spec_path":"docs/planning/specs/4bb367b9_9d6_spec.md","provider":"any","payload_json":"{\"requirements\": {\"coding\": 7, \"reasoning\": 6}, \"_reset_note\": \"This task was reset after a database incident on 2026-04-17.\\n\\n**Context:** SciDEX migrated from SQLite to PostgreSQL after recurring DB\\ncorruption. Some work done during Apr 16-17 may have been lost.\\n\\n**Before starting work:**\\n1. Check if the task's goal is ALREADY satisfied (run the relevant checks)\\n2. Check `git log --all --grep=task:YOUR_TASK_ID` for prior commits\\n3. If complete, verify and mark done. If partial, continue. If not done, proceed.\\n\\n**DB change:** SciDEX now uses PostgreSQL. `get_db()` auto-detects via\\nSCIDEX_DB_BACKEND=postgres env var.\", \"_reset_at\": \"2026-04-18T06:29:22.046013+00:00\", \"_reset_from_status\": \"done\"}"},{"id":"t-tier-classification","title":"Epistemic tier classification engine","description":"epistemic_tiers.py: classify_hypothesis(), classify_edge(), can_override() with asymmetric ratchet logic. Tiers: T0 axiom → T5 contested. Promotion requires evidence, demotion requires stronger counter-evidence.\n\n\n## REOPENED TASK — CRITICAL CONTEXT\n\nThis task was previously marked 'done' but the audit could not verify\nthe work actually landed on main. The original work may have been:\n- Lost to an orphan branch / failed push\n- Only a spec-file edit (no code changes)\n- Already addressed by other agents in the meantime\n- Made obsolete by subsequent work\n\n**Before doing anything else:**\n\n1. **Re-evaluate the task in light of CURRENT main state.** Read the\n spec and the relevant files on origin/main NOW. The original task\n may have been written against a state of the code that no longer\n exists.\n\n2. **Verify the task still advances SciDEX's aims.** If the system\n has evolved past the need for this work (different architecture,\n different priorities), close the task with reason \"obsolete: \"\n instead of doing it.\n\n3. **Check if it's already done.** Run `git log --grep=''`\n and read the related commits. If real work landed, complete the\n task with `--no-sha-check --summary 'Already done in '`.\n\n4. **Make sure your changes don't regress recent functionality.** Many\n agents have been working on this codebase. Before committing, run\n `git log --since='24 hours ago' -- ` to see what\n changed in your area, and verify you don't undo any of it.\n\n5. **Stay scoped.** Only do what this specific task asks for. Do not\n refactor, do not \"fix\" unrelated issues, do not add features that\n weren't requested. Scope creep at this point is regression risk.\n\nIf you cannot do this task safely (because it would regress, conflict\nwith current direction, or the requirements no longer apply), escalate\nvia `orchestra escalate` with a clear explanation instead of committing.\n","status":"done","priority":80,"task_type":"one_shot","frequency":"","assigned_slot":"","started_at":null,"completed_at":"2026-04-19T05:44:34.288372+00:00","updated_at":"2026-04-19T05:44:34.288372+00:00","summary":"","completion_notes":"","last_error":"","time_estimate_hours":0.0,"completion_count":0,"spec_path":"","provider":"any","payload_json":"{\"_reset_note\": \"This task was reset after a database incident on 2026-04-17.\\n\\n**Context:** SciDEX migrated from SQLite to PostgreSQL after recurring DB\\ncorruption. Some work done during Apr 16-17 may have been lost.\\n\\n**Before starting work:**\\n1. Check if the task's goal is ALREADY satisfied (run the relevant checks)\\n2. Check `git log --all --grep=task:YOUR_TASK_ID` for prior commits\\n3. If complete, verify and mark done. If partial, continue. If not done, proceed.\\n\\n**DB change:** SciDEX now uses PostgreSQL. `get_db()` auto-detects via\\nSCIDEX_DB_BACKEND=postgres env var.\", \"_reset_at\": \"2026-04-18T06:29:22.046013+00:00\", \"_reset_from_status\": \"done\"}"},{"id":"t-tier-seeding","title":"Seed epistemic tiers on existing data","description":"Run run_tier_audit() on all 199 hypotheses. Result: 57 established, 26 supported, 66 provisional, 50 contested. Health score: 0.507.\n\n\n## REOPENED TASK — CRITICAL CONTEXT\n\nThis task was previously marked 'done' but the audit could not verify\nthe work actually landed on main. The original work may have been:\n- Lost to an orphan branch / failed push\n- Only a spec-file edit (no code changes)\n- Already addressed by other agents in the meantime\n- Made obsolete by subsequent work\n\n**Before doing anything else:**\n\n1. **Re-evaluate the task in light of CURRENT main state.** Read the\n spec and the relevant files on origin/main NOW. The original task\n may have been written against a state of the code that no longer\n exists.\n\n2. **Verify the task still advances SciDEX's aims.** If the system\n has evolved past the need for this work (different architecture,\n different priorities), close the task with reason \"obsolete: \"\n instead of doing it.\n\n3. **Check if it's already done.** Run `git log --grep=''`\n and read the related commits. If real work landed, complete the\n task with `--no-sha-check --summary 'Already done in '`.\n\n4. **Make sure your changes don't regress recent functionality.** Many\n agents have been working on this codebase. Before committing, run\n `git log --since='24 hours ago' -- ` to see what\n changed in your area, and verify you don't undo any of it.\n\n5. **Stay scoped.** Only do what this specific task asks for. Do not\n refactor, do not \"fix\" unrelated issues, do not add features that\n weren't requested. Scope creep at this point is regression risk.\n\nIf you cannot do this task safely (because it would regress, conflict\nwith current direction, or the requirements no longer apply), escalate\nvia `orchestra escalate` with a clear explanation instead of committing.\n","status":"done","priority":75,"task_type":"one_shot","frequency":"","assigned_slot":"","started_at":null,"completed_at":"2026-04-19T11:15:03.085944+00:00","updated_at":"2026-04-19T11:15:03.085944+00:00","summary":"","completion_notes":"","last_error":"","time_estimate_hours":0.0,"completion_count":0,"spec_path":"","provider":"any","payload_json":"{\"_reset_note\": \"This task was reset after a database incident on 2026-04-17.\\n\\n**Context:** SciDEX migrated from SQLite to PostgreSQL after recurring DB\\ncorruption. Some work done during Apr 16-17 may have been lost.\\n\\n**Before starting work:**\\n1. Check if the task's goal is ALREADY satisfied (run the relevant checks)\\n2. Check `git log --all --grep=task:YOUR_TASK_ID` for prior commits\\n3. If complete, verify and mark done. If partial, continue. If not done, proceed.\\n\\n**DB change:** SciDEX now uses PostgreSQL. `get_db()` auto-detects via\\nSCIDEX_DB_BACKEND=postgres env var.\", \"_reset_at\": \"2026-04-18T06:29:22.046013+00:00\", \"_reset_from_status\": \"done\"}"},{"id":"t-contradiction-check","title":"Contradiction checking for KG edges","description":"contradiction_check() in epistemic_tiers.py — checks if proposed edge conflicts with T0/T1/T2 edges. API: /api/contradiction-check.\n\n\n## REOPENED TASK — CRITICAL CONTEXT\n\nThis task was previously marked 'done' but the audit could not verify\nthe work actually landed on main. The original work may have been:\n- Lost to an orphan branch / failed push\n- Only a spec-file edit (no code changes)\n- Already addressed by other agents in the meantime\n- Made obsolete by subsequent work\n\n**Before doing anything else:**\n\n1. **Re-evaluate the task in light of CURRENT main state.** Read the\n spec and the relevant files on origin/main NOW. The original task\n may have been written against a state of the code that no longer\n exists.\n\n2. **Verify the task still advances SciDEX's aims.** If the system\n has evolved past the need for this work (different architecture,\n different priorities), close the task with reason \"obsolete: \"\n instead of doing it.\n\n3. **Check if it's already done.** Run `git log --grep=''`\n and read the related commits. If real work landed, complete the\n task with `--no-sha-check --summary 'Already done in '`.\n\n4. **Make sure your changes don't regress recent functionality.** Many\n agents have been working on this codebase. Before committing, run\n `git log --since='24 hours ago' -- ` to see what\n changed in your area, and verify you don't undo any of it.\n\n5. **Stay scoped.** Only do what this specific task asks for. Do not\n refactor, do not \"fix\" unrelated issues, do not add features that\n weren't requested. Scope creep at this point is regression risk.\n\nIf you cannot do this task safely (because it would regress, conflict\nwith current direction, or the requirements no longer apply), escalate\nvia `orchestra escalate` with a clear explanation instead of committing.\n","status":"done","priority":70,"task_type":"one_shot","frequency":"","assigned_slot":"","started_at":null,"completed_at":"2026-04-20T16:31:55.700892+00:00","updated_at":"2026-04-20T16:31:55.700892+00:00","summary":"","completion_notes":"Auto-completed by supervisor after successful deploy to main","last_error":"","time_estimate_hours":0.0,"completion_count":0,"spec_path":"","provider":"any","payload_json":"{\"_reset_note\": \"This task was reset after a database incident on 2026-04-17.\\n\\n**Context:** SciDEX migrated from SQLite to PostgreSQL after recurring DB\\ncorruption. Some work done during Apr 16-17 may have been lost.\\n\\n**Before starting work:**\\n1. Check if the task's goal is ALREADY satisfied (run the relevant checks)\\n2. Check `git log --all --grep=task:YOUR_TASK_ID` for prior commits\\n3. If complete, verify and mark done. If partial, continue. If not done, proceed.\\n\\n**DB change:** SciDEX now uses PostgreSQL. `get_db()` auto-detects via\\nSCIDEX_DB_BACKEND=postgres env var.\", \"_reset_at\": \"2026-04-18T06:29:22.046013+00:00\", \"_reset_from_status\": \"done\"}"},{"id":"t-epistemic-dashboard","title":"Epistemic health dashboard","description":"/senate/epistemic-health page with tier distribution bars, replication status, falsification results, health score. API: /api/epistemic-health.\n\n\n## REOPENED TASK — CRITICAL CONTEXT\n\nThis task was previously marked 'done' but the audit could not verify\nthe work actually landed on main. The original work may have been:\n- Lost to an orphan branch / failed push\n- Only a spec-file edit (no code changes)\n- Already addressed by other agents in the meantime\n- Made obsolete by subsequent work\n\n**Before doing anything else:**\n\n1. **Re-evaluate the task in light of CURRENT main state.** Read the\n spec and the relevant files on origin/main NOW. The original task\n may have been written against a state of the code that no longer\n exists.\n\n2. **Verify the task still advances SciDEX's aims.** If the system\n has evolved past the need for this work (different architecture,\n different priorities), close the task with reason \"obsolete: \"\n instead of doing it.\n\n3. **Check if it's already done.** Run `git log --grep=''`\n and read the related commits. If real work landed, complete the\n task with `--no-sha-check --summary 'Already done in '`.\n\n4. **Make sure your changes don't regress recent functionality.** Many\n agents have been working on this codebase. Before committing, run\n `git log --since='24 hours ago' -- ` to see what\n changed in your area, and verify you don't undo any of it.\n\n5. **Stay scoped.** Only do what this specific task asks for. Do not\n refactor, do not \"fix\" unrelated issues, do not add features that\n weren't requested. Scope creep at this point is regression risk.\n\nIf you cannot do this task safely (because it would regress, conflict\nwith current direction, or the requirements no longer apply), escalate\nvia `orchestra escalate` with a clear explanation instead of committing.\n","status":"done","priority":65,"task_type":"one_shot","frequency":"","assigned_slot":"","started_at":null,"completed_at":"2026-04-20T20:37:28.263194+00:00","updated_at":"2026-04-20T20:37:28.263194+00:00","summary":"","completion_notes":"Auto-completed by supervisor after successful deploy to main","last_error":"","time_estimate_hours":0.0,"completion_count":0,"spec_path":"","provider":"any","payload_json":"{\"_reset_note\": \"This task was reset after a database incident on 2026-04-17.\\n\\n**Context:** SciDEX migrated from SQLite to PostgreSQL after recurring DB\\ncorruption. Some work done during Apr 16-17 may have been lost.\\n\\n**Before starting work:**\\n1. Check if the task's goal is ALREADY satisfied (run the relevant checks)\\n2. Check `git log --all --grep=task:YOUR_TASK_ID` for prior commits\\n3. If complete, verify and mark done. If partial, continue. If not done, proceed.\\n\\n**DB change:** SciDEX now uses PostgreSQL. `get_db()` auto-detects via\\nSCIDEX_DB_BACKEND=postgres env var.\", \"_reset_at\": \"2026-04-18T06:29:22.046013+00:00\", \"_reset_from_status\": \"done\"}"},{"id":"t-expiry-system","title":"Auto-expiry for speculative and provisional claims","description":"check_expiry(): speculative items > 30 days without promotion → archived. Provisional items > 90 days without evidence → demoted to speculative.\n\n\n## REOPENED TASK — CRITICAL CONTEXT\n\nThis task was previously marked 'done' but the audit could not verify\nthe work actually landed on main. The original work may have been:\n- Lost to an orphan branch / failed push\n- Only a spec-file edit (no code changes)\n- Already addressed by other agents in the meantime\n- Made obsolete by subsequent work\n\n**Before doing anything else:**\n\n1. **Re-evaluate the task in light of CURRENT main state.** Read the\n spec and the relevant files on origin/main NOW. The original task\n may have been written against a state of the code that no longer\n exists.\n\n2. **Verify the task still advances SciDEX's aims.** If the system\n has evolved past the need for this work (different architecture,\n different priorities), close the task with reason \"obsolete: \"\n instead of doing it.\n\n3. **Check if it's already done.** Run `git log --grep=''`\n and read the related commits. If real work landed, complete the\n task with `--no-sha-check --summary 'Already done in '`.\n\n4. **Make sure your changes don't regress recent functionality.** Many\n agents have been working on this codebase. Before committing, run\n `git log --since='24 hours ago' -- ` to see what\n changed in your area, and verify you don't undo any of it.\n\n5. **Stay scoped.** Only do what this specific task asks for. Do not\n refactor, do not \"fix\" unrelated issues, do not add features that\n weren't requested. Scope creep at this point is regression risk.\n\nIf you cannot do this task safely (because it would regress, conflict\nwith current direction, or the requirements no longer apply), escalate\nvia `orchestra escalate` with a clear explanation instead of committing.\n","status":"done","priority":60,"task_type":"one_shot","frequency":"","assigned_slot":"","started_at":null,"completed_at":"2026-04-20T20:24:46.903134+00:00","updated_at":"2026-04-20T20:24:46.903134+00:00","summary":"","completion_notes":"","last_error":"","time_estimate_hours":0.0,"completion_count":0,"spec_path":"","provider":"any","payload_json":"{\"_reset_note\": \"This task was reset after a database incident on 2026-04-17.\\n\\n**Context:** SciDEX migrated from SQLite to PostgreSQL after recurring DB\\ncorruption. Some work done during Apr 16-17 may have been lost.\\n\\n**Before starting work:**\\n1. Check if the task's goal is ALREADY satisfied (run the relevant checks)\\n2. Check `git log --all --grep=task:YOUR_TASK_ID` for prior commits\\n3. If complete, verify and mark done. If partial, continue. If not done, proceed.\\n\\n**DB change:** SciDEX now uses PostgreSQL. `get_db()` auto-detects via\\nSCIDEX_DB_BACKEND=postgres env var.\", \"_reset_at\": \"2026-04-18T06:29:22.046013+00:00\", \"_reset_from_status\": \"done\"}"}],"reviews":[],"effectiveness":{},"spec_content":"---\ntitle: \"Epistemic Rigor — Falsifiable Hypotheses, Evidence Chains, and Trust Propagation\"\ndescription: \"Quest to evolve SciDEX toward rigorous scientific epistemology\"\nstatus: active\npriority: 95\nlayer: Cross-cutting\nquest: q-epistemic-rigor\n---\n\n# Quest: Epistemic Rigor\n\n## Vision\n\nEvery claim in SciDEX should be **falsifiable, traceable, versioned, and trust-scored**.\n\nToday, hypotheses are scored by composite metrics but lack explicit testable predictions.\nEvidence is stored as JSON blobs without provenance. Knowledge graph edges have no trust scores.\nThere's no dependency structure between hypotheses, experiments, and evidence. Score changes\nhappen without structured justification.\n\nThis quest transforms SciDEX from \"we scored this hypothesis 0.73\" to \"this hypothesis\npredicts X, experiment Y tested it, the result was Z, which updated our confidence because\nof evidence chain A->B->C, each link traceable to ground truth with trust score T.\"\n\n## Current State (What Exists)\n\n| Component | Status | Gap |\n|-----------|--------|-----|\n| 10-dimension hypothesis scoring | Working | No explicit predictions or falsification criteria |\n| Evidence for/against (JSON) | Working | Unstructured, no provenance, no methodology |\n| Evidence validation (PMID relevance) | Working | Scores relevance, not trustworthiness |\n| Belief snapshots (time series) | Working | Tracks score evolution but not WHY scores changed |\n| Debate quality scoring | Working | Includes falsifiability dimension, but not structured |\n| Persona believability | Working | Per-dimension credibility, but no update from outcomes |\n| Experiments table | Working | No results storage, no prediction-vs-reality comparison |\n| Knowledge graph edges | Working | evidence_strength field but no trust model |\n| Price history with events | Working | event_source is free text, not structured provenance |\n| Quality gates | Working | Code quality, not epistemic quality |\n\n## Architecture: 8 Tasks in Dependency Order\n\n```\nTask 1: Predictions Table ──┬──> Task 2: Experiment Results\n │\nTask 3: Evidence Chains ────┼──> Task 4: Trust Scores on KG\n │\n ├──> Task 5: Dependency Graph\n │\nTask 6: Versioning/Audit ───┘\n \nTask 7: Knowledge Units (depends on 3, 4, 6)\n\nTask 8: Epistemic Dashboard (depends on all above)\n```\n\n## Key Design Principles\n\n1. **Ground truth anchoring**: Every claim must trace to a paper, experiment, or dataset\n2. **Bayesian updating**: New evidence should update confidence via structured reasoning, not just recompute\n3. **Falsifiability first**: Hypotheses without predictions are speculation, not science\n4. **Trust propagation**: Downstream conclusions inherit (diminished) trust from upstream evidence\n5. **Audit completeness**: Every score change has a structured justification\n6. **Composability**: Evidence blocks are atomic, addressable, and combinable\n7. **Incremental delivery**: Each task is independently valuable and deployable\n8. **Experiment-boosted ranking**: Hypotheses with explicit falsifiable predictions AND high-quality associated experiments (feasible, impactful) should receive a composite score boost. The system should reward hypotheses that are not just well-scored but actively testable with concrete, feasible experiments.\n\n## Hypothesis-Experiment Scoring Feedback Loop\n\nThe composite score formula should incorporate a **testability bonus** that rewards hypotheses linked to high-quality experiments:\n\n### Scoring Considerations\n\n1. **Falsifiability bonus**: Hypotheses with explicit, testable predictions (via `hypothesis_predictions` table) receive a score multiplier. A hypothesis that merely claims \"X causes Y\" ranks lower than one that predicts \"If X, then Y should be measurable as Z with effect size > threshold.\"\n\n2. **Experiment quality signal**: When a hypothesis has associated experiments (via `experiments.hypothesis_ids`), the experiment's own quality scores feed back into the hypothesis ranking:\n - **Feasibility**: A hypothesis testable by a practical, affordable experiment is more valuable than one requiring impossible resources\n - **Impact**: A hypothesis whose experiment would significantly update the world model (high information gain) ranks higher\n - **Experiment composite** = feasibility × 0.4 + impact × 0.4 + novelty × 0.2\n\n3. **Combined boost formula**:\n ```\n testability_bonus = 0.0\n if has_falsifiable_predictions:\n testability_bonus += 0.05\n if has_associated_experiments:\n avg_experiment_quality = mean(exp.composite for exp in linked_experiments)\n testability_bonus += 0.10 * avg_experiment_quality\n adjusted_composite = base_composite + testability_bonus\n ```\n\n4. **Virtuous cycle**: This creates a feedback loop where:\n - Hypotheses with predictions attract experiment design\n - High-quality experiments boost hypothesis ranking\n - Higher-ranked hypotheses get more attention and resources\n - More attention produces better predictions and experiments\n\n### Implementation Notes\n\n- The testability bonus should be computed in `post_process.py` alongside the existing composite score\n- Requires Task 1 (predictions table) and Task 2 (experiment results) to be complete\n- The bonus is additive, not multiplicative, to avoid runaway scores\n- Experiments without results still provide a feasibility/impact signal\n- Cap the total bonus at 0.15 to prevent gaming\n\n## Key Files (Existing)\n\n- `exchange.py` — 10-dim scoring, believability weighting, allocation\n- `senate_proposals.py` — Evidence strength scoring (papers + citations + recency)\n- `evidence_validator.py` — PMID relevance scoring via Claude Haiku\n- `belief_tracker.py` — Temporal belief snapshots, convergence metrics\n- `backfill_debate_quality.py` — 4-dim debate quality (includes falsifiability)\n- `quality_gates.py` — Pre-merge, post-completion, prevention gates\n- `market_dynamics.py` — LMSR price model with event logging\n\n## Key Tables (Existing)\n\n- `hypotheses` — evidence_for/against (JSON), evidence_validation_score\n- `hypothesis_papers` — Junction: hypothesis <-> PMID with direction/claim/strength\n- `papers` — PubMed metadata (pmid, title, abstract, citations, year)\n- `knowledge_edges` — source/target with evidence_strength (REAL)\n- `experiments` — hypothesis_ids (JSON), protocol, expected_outcomes, status\n- `debate_sessions` — transcript_json, quality_score\n- `debate_rounds` — evidence_cited (JSON), hypotheses_referenced\n- `belief_snapshots` — Time series of hypothesis score + evidence count\n- `price_history` — Score changes with event_type and event_source\n- `persona_believability` — Per-persona, per-dimension credibility\n- `edit_history` — Generic audit log (actor, content_type, diff, reason)\n\n## Workstreams\n\n### WS-rigor-ruleset — Absorb Alpha1 Science's 8-dimension biomedical rigor rubric\n\nAbsorb Alpha1 Science's 8-dimension rigor rubric (scientific premise,\nstudy design, blinding, power analysis, resource identification,\nstatistical reporting, data availability + 1 TBD) grounded in\n**NIH / MDAR / ARRIVE 2.0 / CONSORT / EQUATOR** biomedical reporting\nguidelines. Every hypothesis and analysis gets a rigor score card.\nEvery score carries an **evidence citation** pointing to the exact\ntext it was derived from. Score card becomes a new artifact type;\noptionally publishable to the SciDEX community view. See\n[`docs/bio_competitive/alpha1_science_profile.md`](../../bio_competitive/alpha1_science_profile.md).\n\n**Deliverables:**\n- Rubric dictionary: JSON mapping NIH / MDAR / ARRIVE 2.0 / CONSORT /\n EQUATOR items to the 8 dimensions, with specific guideline-item\n pointers per dimension.\n- Two-agent independent evaluation pipeline. Reuse the Skeptic\n persona; the second agent is a separately-seeded Skeptic instance\n (different prompt seed or different provider) to preserve\n independence — matches Alpha1's 2-independent-agent design.\n- Rigor score card as a first-class artifact type in the `artifacts`\n table, with lineage to the hypothesis or analysis it scores.\n- Evidence-citation schema: every score row carries a\n `source_quote` + `source_location` (PMID / page / paragraph) so a\n reviewer can verify the score without re-reading the whole paper.\n- Community-publish surface: optional Atlas wiki page per score card\n for the ones authors opt to publish.\n- Recurring Senate-layer task to produce score cards for the\n backlog; see\n [`docs/planning/specs/task-id-pending_rigor_score_card_spec.md`](task-id-pending_rigor_score_card_spec.md).\n\n**Dependency on existing tasks:**\n- Task 3 (Evidence Chains) provides the structured-evidence table\n the rigor score card cites.\n- Task 4 (Trust Scores on KG) consumes the rigor score to weight KG\n edges drawn from papers with a published rigor score.\n- Task 7 (Knowledge Units) can include a rigor-score summary as one\n of its atomic evidence blocks.\n\n**Success metric:** every hypothesis and every ≥50KB analysis in the\ntrailing 30 days has a rigor score card with ≥95% of scores carrying\nan evidence citation; inter-rater agreement between the two\nindependent agents ≥0.7 (Cohen's κ or equivalent).\n\n## Related Quests\n\n| Quest | Relationship |\n|-------|-------------|\n| **Experiment Extraction** (q-experiment-extraction) | Structured experiments are the ground-truth anchors for evidence chains |\n| **Artifact Debates** (q-artifact-debates) | Any artifact can be debated, accumulating evidence about its quality |\n| **Schema Governance** (q-schema-governance) | Evidence schemas evolve through governance to maintain integrity |\n| **Artifacts** (8db4834c-51e) | All evidence is stored as artifacts with lineage and provenance |\n| **Competitive Biotools** (q-competitive-biotools) | Tracks Alpha1 Science + PRISM; the WS-rigor-ruleset workstream absorbs Alpha1's 8-dim rubric into this quest |\n\n### How These Quests Interlock\n\nThe epistemic rigor vision depends on three supporting quests:\n\n1. **Experiment Extraction** provides the **ground truth** — structured experimental results\n with p-values, effect sizes, and methodology that anchor evidence chains to reality.\n\n2. **Artifact Debates** provides the **self-correction mechanism** — when evidence is\n contested, structured debates accumulate arguments for and against, and quality scores\n update based on debate outcomes.\n\n3. **Schema Governance** provides the **integrity guarantee** — as we learn what evidence\n structure is useful, schemas evolve through agent governance rather than ad-hoc changes,\n ensuring data remains queryable and trustworthy.\n\nTogether: experiments ground claims in data, debates correct errors, governance maintains integrity.\n\n## Success Metrics\n\n- **Falsifiability coverage**: >80% of hypotheses have explicit testable predictions\n- **Provenance coverage**: >90% of evidence claims traced to source (paper/experiment/debate)\n- **Trust score coverage**: 100% of KG edges have computed trust scores\n- **Audit completeness**: 100% of score changes have structured justifications\n- **Dependency graph**: All hypothesis-experiment links are bidirectional and queryable\n- **Experiment grounding**: >500 structured experiments extracted, each traceable to source paper\n- **Debate breadth**: >5 artifact types have been debated (not just hypotheses)\n- **Schema integrity**: All artifact types have governed schemas with validation\n\n## Code Quality Requirements\n\nAll code produced by this quest must:\n- Use shared `database.py` for DB connections (not local `get_db()` definitions)\n- Include migration testing (`--dry-run` verification before apply)\n- Add tests for trust computation, Bayesian score updates, and provenance chain traversal\n- New modules must be < 500 lines; split if larger\n- No duplicate utility functions — reuse from `pubmed_utils.py`, `kg_extraction_utils.py`\n- Schema changes reviewed for index coverage and query performance\n","spec_html":"

Quest: Epistemic Rigor

Vision

Every claim in SciDEX should be falsifiable, traceable, versioned, and trust-scored.

Today, hypotheses are scored by composite metrics but lack explicit testable predictions.
Evidence is stored as JSON blobs without provenance. Knowledge graph edges have no trust scores.
There's no dependency structure between hypotheses, experiments, and evidence. Score changes
happen without structured justification.

This quest transforms SciDEX from "we scored this hypothesis 0.73" to "this hypothesis
predicts X, experiment Y tested it, the result was Z, which updated our confidence because
of evidence chain A->B->C, each link traceable to ground truth with trust score T."

Current State (What Exists)

ComponentStatusGap
10-dimension hypothesis scoringWorkingNo explicit predictions or falsification criteria
Evidence for/against (JSON)WorkingUnstructured, no provenance, no methodology
Evidence validation (PMID relevance)WorkingScores relevance, not trustworthiness
Belief snapshots (time series)WorkingTracks score evolution but not WHY scores changed
Debate quality scoringWorkingIncludes falsifiability dimension, but not structured
Persona believabilityWorkingPer-dimension credibility, but no update from outcomes
Experiments tableWorkingNo results storage, no prediction-vs-reality comparison
Knowledge graph edgesWorkingevidence_strength field but no trust model
Price history with eventsWorkingevent_source is free text, not structured provenance
Quality gatesWorkingCode quality, not epistemic quality
\n

Architecture: 8 Tasks in Dependency Order

Task 1: Predictions Table ──┬──> Task 2: Experiment Results\n                            │\nTask 3: Evidence Chains ────┼──> Task 4: Trust Scores on KG\n                            │\n                            ├──> Task 5: Dependency Graph\n                            │\nTask 6: Versioning/Audit ───┘\n                            \nTask 7: Knowledge Units (depends on 3, 4, 6)\n\nTask 8: Epistemic Dashboard (depends on all above)

Key Design Principles

  • Ground truth anchoring: Every claim must trace to a paper, experiment, or dataset
    • \n
  • Bayesian updating: New evidence should update confidence via structured reasoning, not just recompute
    • \n
  • Falsifiability first: Hypotheses without predictions are speculation, not science
    • \n
  • Trust propagation: Downstream conclusions inherit (diminished) trust from upstream evidence
    • \n
  • Audit completeness: Every score change has a structured justification
    • \n
  • Composability: Evidence blocks are atomic, addressable, and combinable
    • \n
  • Incremental delivery: Each task is independently valuable and deployable
    • \n
  • Experiment-boosted ranking: Hypotheses with explicit falsifiable predictions AND high-quality associated experiments (feasible, impactful) should receive a composite score boost. The system should reward hypotheses that are not just well-scored but actively testable with concrete, feasible experiments.

    • Hypothesis-Experiment Scoring Feedback Loop

    The composite score formula should incorporate a testability bonus that rewards hypotheses linked to high-quality experiments:

    Scoring Considerations

  • Falsifiability bonus: Hypotheses with explicit, testable predictions (via hypothesis_predictions table) receive a score multiplier. A hypothesis that merely claims "X causes Y" ranks lower than one that predicts "If X, then Y should be measurable as Z with effect size > threshold."

  • Experiment quality signal: When a hypothesis has associated experiments (via experiments.hypothesis_ids), the experiment's own quality scores feed back into the hypothesis ranking:
    • \n - Feasibility: A hypothesis testable by a practical, affordable experiment is more valuable than one requiring impossible resources
    - Impact: A hypothesis whose experiment would significantly update the world model (high information gain) ranks higher
    - Experiment composite = feasibility × 0.4 + impact × 0.4 + novelty × 0.2

  • Combined boost formula:
    • \n 
    testability_bonus = 0.0\n   if has_falsifiable_predictions:\n       testability_bonus += 0.05\n   if has_associated_experiments:\n       avg_experiment_quality = mean(exp.composite for exp in linked_experiments)\n       testability_bonus += 0.10 * avg_experiment_quality\n   adjusted_composite = base_composite + testability_bonus

  • Virtuous cycle: This creates a feedback loop where:
    • \n - Hypotheses with predictions attract experiment design
    - High-quality experiments boost hypothesis ranking
    - Higher-ranked hypotheses get more attention and resources
    - More attention produces better predictions and experiments

    Implementation Notes

    • The testability bonus should be computed in post_process.py alongside the existing composite score
      • \n
    • Requires Task 1 (predictions table) and Task 2 (experiment results) to be complete
      • \n
    • The bonus is additive, not multiplicative, to avoid runaway scores
      • \n
    • Experiments without results still provide a feasibility/impact signal
      • \n
    • Cap the total bonus at 0.15 to prevent gaming
      • \n
    \n

    Key Files (Existing)

    • exchange.py — 10-dim scoring, believability weighting, allocation
      • \n
    • senate_proposals.py — Evidence strength scoring (papers + citations + recency)
      • \n
    • evidence_validator.py — PMID relevance scoring via Claude Haiku
      • \n
    • belief_tracker.py — Temporal belief snapshots, convergence metrics
      • \n
    • backfill_debate_quality.py — 4-dim debate quality (includes falsifiability)
      • \n
    • quality_gates.py — Pre-merge, post-completion, prevention gates
      • \n
    • market_dynamics.py — LMSR price model with event logging
      • \n
    \n

    Key Tables (Existing)

    • hypotheses — evidence_for/against (JSON), evidence_validation_score
      • \n
    • hypothesis_papers — Junction: hypothesis <-> PMID with direction/claim/strength
      • \n
    • papers — PubMed metadata (pmid, title, abstract, citations, year)
      • \n
    • knowledge_edges — source/target with evidence_strength (REAL)
      • \n
    • experiments — hypothesis_ids (JSON), protocol, expected_outcomes, status
      • \n
    • debate_sessions — transcript_json, quality_score
      • \n
    • debate_rounds — evidence_cited (JSON), hypotheses_referenced
      • \n
    • belief_snapshots — Time series of hypothesis score + evidence count
      • \n
    • price_history — Score changes with event_type and event_source
      • \n
    • persona_believability — Per-persona, per-dimension credibility
      • \n
    • edit_history — Generic audit log (actor, content_type, diff, reason)
      • \n
    \n

    Workstreams

    WS-rigor-ruleset — Absorb Alpha1 Science's 8-dimension biomedical rigor rubric

    Absorb Alpha1 Science's 8-dimension rigor rubric (scientific premise,
    study design, blinding, power analysis, resource identification,
    statistical reporting, data availability + 1 TBD) grounded in\nNIH / MDAR / ARRIVE 2.0 / CONSORT / EQUATOR biomedical reporting
    guidelines. Every hypothesis and analysis gets a rigor score card.
    Every score carries an evidence citation pointing to the exact
    text it was derived from. Score card becomes a new artifact type;
    optionally publishable to the SciDEX community view. See
    [docs/bio_competitive/alpha1_science_profile.md](../../bio_competitive/alpha1_science_profile.md).

    Deliverables:\n

    • Rubric dictionary: JSON mapping NIH / MDAR / ARRIVE 2.0 / CONSORT /
      • \n
    EQUATOR items to the 8 dimensions, with specific guideline-item
    pointers per dimension.\n
    • Two-agent independent evaluation pipeline. Reuse the Skeptic
      • \n
    persona; the second agent is a separately-seeded Skeptic instance
    (different prompt seed or different provider) to preserve
    independence — matches Alpha1's 2-independent-agent design.\n
    • Rigor score card as a first-class artifact type in the artifacts
      • \n
    table, with lineage to the hypothesis or analysis it scores.\n
    • Evidence-citation schema: every score row carries a
      • \n
    source_quote + source_location (PMID / page / paragraph) so a
    reviewer can verify the score without re-reading the whole paper.\n
    • Community-publish surface: optional Atlas wiki page per score card
      • \n
    for the ones authors opt to publish.\n
    • Recurring Senate-layer task to produce score cards for the
      • \n
    backlog; see
    [docs/planning/specs/task-id-pending_rigor_score_card_spec.md](task-id-pending_rigor_score_card_spec.md).

    Dependency on existing tasks:\n

    • Task 3 (Evidence Chains) provides the structured-evidence table
      • \n
    the rigor score card cites.\n
    • Task 4 (Trust Scores on KG) consumes the rigor score to weight KG
      • \n
    edges drawn from papers with a published rigor score.\n
    • Task 7 (Knowledge Units) can include a rigor-score summary as one
      • \n
    of its atomic evidence blocks.

    Success metric: every hypothesis and every ≥50KB analysis in the
    trailing 30 days has a rigor score card with ≥95% of scores carrying
    an evidence citation; inter-rater agreement between the two
    independent agents ≥0.7 (Cohen's κ or equivalent).

    Related Quests

    QuestRelationship
    Experiment Extraction (q-experiment-extraction)Structured experiments are the ground-truth anchors for evidence chains
    Artifact Debates (q-artifact-debates)Any artifact can be debated, accumulating evidence about its quality
    Schema Governance (q-schema-governance)Evidence schemas evolve through governance to maintain integrity
    Artifacts (8db4834c-51e)All evidence is stored as artifacts with lineage and provenance
    Competitive Biotools (q-competitive-biotools)Tracks Alpha1 Science + PRISM; the WS-rigor-ruleset workstream absorbs Alpha1's 8-dim rubric into this quest
    \n

    How These Quests Interlock

    The epistemic rigor vision depends on three supporting quests:

  • Experiment Extraction provides the ground truth — structured experimental results
    • \n with p-values, effect sizes, and methodology that anchor evidence chains to reality.

  • Artifact Debates provides the self-correction mechanism — when evidence is
    • \n contested, structured debates accumulate arguments for and against, and quality scores
    update based on debate outcomes.

  • Schema Governance provides the integrity guarantee — as we learn what evidence
    • \n structure is useful, schemas evolve through agent governance rather than ad-hoc changes,
    ensuring data remains queryable and trustworthy.

    Together: experiments ground claims in data, debates correct errors, governance maintains integrity.

    Success Metrics

    • Falsifiability coverage: >80% of hypotheses have explicit testable predictions
      • \n
    • Provenance coverage: >90% of evidence claims traced to source (paper/experiment/debate)
      • \n
    • Trust score coverage: 100% of KG edges have computed trust scores
      • \n
    • Audit completeness: 100% of score changes have structured justifications
      • \n
    • Dependency graph: All hypothesis-experiment links are bidirectional and queryable
      • \n
    • Experiment grounding: >500 structured experiments extracted, each traceable to source paper
      • \n
    • Debate breadth: >5 artifact types have been debated (not just hypotheses)
      • \n
    • Schema integrity: All artifact types have governed schemas with validation
      • \n
    \n

    Code Quality Requirements

    All code produced by this quest must:\n

    • Use shared database.py for DB connections (not local get_db() definitions)
      • \n
    • Include migration testing (--dry-run verification before apply)
      • \n
    • Add tests for trust computation, Bayesian score updates, and provenance chain traversal
      • \n
    • New modules must be < 500 lines; split if larger
      • \n
    • No duplicate utility functions — reuse from pubmed_utils.py, kg_extraction_utils.py
      • \n
    • Schema changes reviewed for index coverage and query performance
      • \n

    ","spec_file":"quest_epistemic_rigor.md","commits":[{"hash":"59405c7c5","message":"docs: AGENTS.md — document Path A/B/C task completion semantics [task:docs-agents-completion] (#40)","date":"2026-04-25"},{"hash":"e5b5848a0","message":"WIP on orchestra/task/8fcc8dc8-debate-artifact-version-pinning-referenc: 8a24c2fa2 [Senate] Delete broken restore_database.sh (#38)","date":"2026-04-25"},{"hash":"50e5ffcfe","message":"index on orchestra/task/8fcc8dc8-debate-artifact-version-pinning-referenc: 8a24c2fa2 [Senate] Delete broken restore_database.sh (#38)","date":"2026-04-25"},{"hash":"0d37f5fce","message":"untracked files on orchestra/task/8fcc8dc8-debate-artifact-version-pinning-referenc: 8a24c2fa2 [Senate] Delete broken restore_database.sh (#38)","date":"2026-04-25"},{"hash":"48f8d2fe3","message":"feat: surface all five SciDEX layers in nav [task:cba19c94-1724-4d5a-b89d-96c73c25f12a] (#39)","date":"2026-04-25"},{"hash":"1f0e35929","message":"Squash merge: orchestra/task/b1a8e549-cross-cutting-wire-existing-k-dense-skil (2 commits)","date":"2026-04-25"},{"hash":"ddb7db381","message":"[Agora] Wire existing K-Dense-backed tools into debate orchestration [task:b1a8e549-6f31-43c5-80f5-7c4717c267e4]","date":"2026-04-25"},{"hash":"76b71427a","message":"[Agora] Wire existing K-Dense-backed tools into debate orchestration [task:b1a8e549-6f31-43c5-80f5-7c4717c267e4]","date":"2026-04-25"},{"hash":"779e85c3a","message":"[Senate] Verify /resources dashboard complete; check off acceptance criteria [task:82074adc-507f-4e6b-9092-e2ceee79e7d4]","date":"2026-04-25"},{"hash":"4c66a8e09","message":"[Senate] Establish emergency access recovery procedures [task:e643cdd3-afd6-410f-a366-a6297d112127]","date":"2026-04-25"},{"hash":"7265a06b4","message":"Squash merge: orchestra/task/b1a8e549-cross-cutting-wire-existing-k-dense-skil (1 commits)","date":"2026-04-25"},{"hash":"58406ec64","message":"[Atlas] Dashboard artifact type: living web views with data source rendering [task:a17-28-DASH0001]","date":"2026-04-25"},{"hash":"8a24c2fa2","message":"[Senate] Delete broken restore_database.sh (#38)","date":"2026-04-25"},{"hash":"b98a1fa18","message":"[Senate] Delete broken restore_database.sh","date":"2026-04-25"},{"hash":"e846f82ef","message":"[Senate] Refresh BACKUP_RESTORE.md + docs/runbooks/emergency_restore.md (#37)","date":"2026-04-25"},{"hash":"43972a45e","message":"[Senate] Refresh BACKUP_RESTORE.md + docs/runbooks/emergency_restore.md","date":"2026-04-25"},{"hash":"2c7dbfe7f","message":"[Senate] Delete 9 obsolete backup scripts/units (continuation of Phase A-D cleanup) (#36)","date":"2026-04-25"},{"hash":"9743eb298","message":"[Senate] Delete 9 obsolete backup scripts/units (continuation of Phase A-D cleanup)","date":"2026-04-25"},{"hash":"3e72d8383","message":"[Agora] Wire 3 missing tools into debate skill_functions, fix citation persistence bug [task:b1a8e549-6f31-43c5-80f5-7c4717c267e4]","date":"2026-04-25"},{"hash":"4310e9854","message":"[Demo] Work log: figures verified complete — 140/140 analyses covered [task:df201d8f-4b89-4258-9148-eb1028fc1fbd]","date":"2026-04-24"}]}