メニュー
Invoke for claims.json final output format reference. Defines the schema for aggregated claims across all investigations.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Invoke after experiment-runner completes an investigation's query bundle. Evaluates all claims in current_investigation.json and produces an investigation-level adjudication. Writes claim_scores.json.
Invoke after investigation-decomposition. Runs all queries in an investigation bundle using shared data preparation. Writes code and results.
Invoke after main agent writes L2 assessment. Reads epistemic state (L1/L2), patterns, and investigations to recommend next action. Writes strategy_recommendation.json.
Gene set enrichment analysis (GSEA, Enrichr, over-representation). Invoke when query mentions enrichment, pathway analysis, GO analysis, GSEA, or gene set.
Invoke before running experiments for an investigation. Reads current_investigation.json and task_packet.json, writes current_investigation_requirements.json with shared data preparation and per-query requirements.
Gene ID conversion and annotation. Invoke when you need to map between any gene identifier formats.
| name | write-claims |
| description | Invoke for claims.json final output format reference. Defines the schema for aggregated claims across all investigations. |
Write to /workspace/run/claims.json. This is the final output file aggregating all claims from all investigations. Updated incrementally after each judge evaluation, finalized at the end of exploration.
| Field | Type | Required | Description |
|---|---|---|---|
task_id | string | yes | From task_packet.json |
claims | array | yes | All claims from all investigations (see below) |
top_supported_claim_id | string | yes | The claim ID of the strongest overall finding. Set at finalization. |
investigation_summary | array | yes | Summary of each investigation's outcome (see below) |
| Field | Type | Required | Description |
|---|---|---|---|
claim_id | string | yes | Claim identifier (e.g. I1-claim1) |
investigation_id | string | yes | Which investigation produced this claim |
investigation_name | string | yes | Short label of the investigation |
role | string | yes | One of: main, alternative_explanation, artifact_check, robustness |
text | string | yes | Claim statement with key finding and numbers |
status | string | yes | One of: supported, weakened, refuted, inconclusive |
confidence | number | yes | Judge's calibrated confidence (0-1) |
dimensions | object | yes | See dimensions below |
evidence_refs | array of strings | yes | Experiment IDs that produced evidence for this claim |
supporting_stats | object | no | Key statistics from the experiments |
| Field | Type | Required | Description |
|---|---|---|---|
variables | array of strings | yes | Column names involved |
relation | string | yes | Type of relationship found |
context | string | yes | Analysis context description |
direction | string | yes | Direction or nature of the finding |
| Field | Type | Required | Description |
|---|---|---|---|
investigation_id | string | yes | Investigation identifier |
name | string | yes | Short label |
phenomenon | string | yes | Full phenomenon description |
status | string | yes | Final investigation status |
winning_claim_id | string or null | yes | The claim that best explains the phenomenon |
claim_outcome_counts | object | yes | Counts by claim status: supported, weakened, refuted, inconclusive |
top_supported_claim_id and complete investigation_summary for all investigations.