Menu
Populate FINDINGS.md with results and analysis, update HYPOTHESIS.md status
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Formulate a single testable hypothesis by scanning the project for untested behaviors, gaps, and claims
Guided hypothesis-driven experimentation. Choose a project, generate hypotheses, select which to test, run experiments in parallel or sequentially, document findings.
Generate iteratively-reviewed research ideas from a problem statement. Fully guided - no arguments required.
Send Claude plan to an LLM for external technical review
Create research document with problem statement and context from multiple sources (repositories, papers, files)
Execute experiment run.sh, run analysis script, parse output, and interpret results
| name | _document-findings |
| description | Populate FINDINGS.md with results and analysis, update HYPOTHESIS.md status |
| user-invocable | false |
| allowed-tools | ["Read","Write","Edit","Glob","Grep"] |
Populate FINDINGS.md with experiment results and analysis. Called by background agents.
Skill(_document-findings "<hypothesis_dir>" "<verdict>" "<summary>" "<metrics_table>" "<analysis_output>")
Arguments:
hypothesis_dir — path to hypothesis directory (e.g., hypotheses/h3-caching-latency)verdict — Confirmed, Refuted, Inconclusive, or Failedsummary — 2-3 sentence interpretationmetrics_table — formatted comparison tableanalysis_output — raw output from analyze scriptRead the existing FINDINGS.md template in the hypothesis directory. Update:
Pending with verdict<!-- Auto-populated --> with metrics comparison table (formatted as markdown table), per-configuration output summaries, key numbers highlighted<!-- Auto-populated --> with verdict, reasoning, effect size, surprises. Use Grep and Read to trace mechanisms in source code — include file:line citations explaining WHY the result occurred.Edit HYPOTHESIS.md in the same directory:
**Status**: Pending with **Status**: <verdict>Do NOT update hypotheses/README.md. The orchestrator handles catalog updates after all agents complete to prevent race conditions in parallel mode.
Return this structured text to the calling skill:
FINDINGS_PATH: hypotheses/h<N>-<slug>/FINDINGS.md
VERDICT: <Confirmed|Refuted|Inconclusive|Failed>