REA CI and Tests Audit
Audits REA's CI workflows and test suites for false passes, fixture drift, and safe speedups
- Repository
- morluto/rea
- Stars
- 420,737
- Licence
- MIT
- Files
- 3 files
What it does
When installed in the REA repository, this skill tells an AI agent how to investigate whether its continuous integration checks and tests still catch real bugs, and whether slow parts can be made faster without weakening that protection. The agent is directed to trace workflow triggers and test discovery, reproduce suspected problems against real behavior, deliberately break code to confirm a test would notice, and measure timing before claiming a speedup. It is also told not to edit code, push, or merge unless the user has explicitly authorized it.
Great for
- Maintainers of the REA project investigating flaky, slow, or suspiciously green CI runs
- Contributors who need an evidence-based audit of test coverage or fixture accuracy before proposing CI changes
Skip it if
- Ordinary feature development or diagnosing a single test failure, which the skill's own description says it should not be used for
- Repositories other than REA, since it depends on REA's AGENTS.md, testing guide and workflow layout; applying it elsewhere would mostly produce confusion
- Unattended CI changes: it can lead to edits, reruns and full local suite runs, so it should be used with a human reviewing scope and any proposed merge
Try asking
- “Audit whether the REA CI workflow selects and runs all the required test files, and report any that could be silently skipped”
- “Check whether the test fixtures for the protocol still match current admission requirements and protocol replies”
- “Find ways to cut CI runtime without dropping any assertions, and report measured before and after timings”
- “Verify that a suspicious always-green test actually fails when the behaviour it covers is broken”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 3 original files, the MIT licence and an attribution note.
rea-ci-tests-audit.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip rea-ci-tests-audit.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip rea-ci-tests-audit.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- State the exact scope up front (for example, audit only the CI workflow, or only one test family) so the agent does not drift into broad full-suite runs or edits
- Ask for raw outputs and baseline SHAs to be kept in a /tmp directory, as the skill recommends, and review them before accepting any conclusion
- Review any proposed change to timeouts, retries or test deletions manually, since the skill itself treats those as needing separate justification
What's inside
- agents/openai.yaml
- references/measurement-and-ci.md
- SKILL.md
Redistributed unmodified from morluto/rea under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.