report-generation-skill

name	report-generation-skill
description	This skill should be used when the user asks to generate a report, summarize results, write up findings, or create documentation for a computational math reproduction or tuning experiment.

Report Generation Skill

Generates compact Markdown reports for the Skill-first Codex workflow.

For end-to-end computational math research-code reproduction workflows, this Skill should be selected by computational_math_reproduction_workflow_skill rather than used as the first entrypoint.

Report Types

Four compact Markdown report types:

1. plan.md

# Reproduction Plan

## Goal

## Selected Source

## Minimal Command To Run

## Expected Outputs

## Risks

## Approval Scope

Reply with approve / revise / reject / skip.

2. repair_plan.md

# Repair Plan

## Failure Summary

## Evidence

## Proposed Minimal Repair

## Files To Modify

## What Will Not Be Changed

## Patch / Rollback Plan

Reply with approve / revise / reject / skip.

3. RUN_SUMMARY.md

# Run Summary

## Status

## Source

## Commands Run

## Evidence

## Results

## Figures

## Patches

## Limitations

## Optional Tuning Recommendation

4. tuning/TUNING_SUMMARY.md

# Tuning Summary

## Status

## Budget

## Search Method

## Best Parameters

## Baseline vs Best

## Evidence

## Limitations

Workflow

1. Draft plan.md before execution when a durable plan is useful.
2. After reproduction, write RUN_SUMMARY.md with status, evidence, limits, and next options.
3. If tuning was approved and run, write tuning/TUNING_SUMMARY.md.
4. Ask the user if conclusions are acceptable or if sections need revision.

Optional Helper

report_writer

Input: --run <path> — run directory with execution logs, optional tuning outputs, and figures.

python -m skills.report_generation_skill.scripts.report_writer --run /path/to/output