fcakyon/claude-codex-settings / plugins/test-audit/skills/test-audit
test-audit
VerifiedThis skill should be used when the user asks to "write a test", "review these tests", "audit tests", "prune low-value tests", or "find duplicate tests", and whenever writing, changing, reviewing, or sweeping tests. Gates new tests and audits low-value, implementation-coupled, or duplicative tests and the test-only production seams they demand.
| File | Kind | Size |
|---|---|---|
| LICENSE | other | 1.1 KB |
| SKILL.md | skill md | 2.8 KB |
| references/audit.md | markdown | 3.3 KB |
| references/campaign.md | markdown | 5.4 KB |
references/audit.md
View on GitHub# Audit workflow
- Use after [authoring gate](../SKILL.md#authoring-gate).
- Whole-subsystem campaigns: also require [CAMPAIGN.md](campaign.md).
## Junk patterns
Reject new matches unless [retention bar](#retention-bar) identifies independent
contract. Audit existing matches:
- Assertion-free probes, self-comparisons, identity copiers, subject-generated
expected values.
- Copied inventories/manifests/export lists; exact source/import/string greps.
- Private predicate/call-shape tests duplicating boundary proof; repeated
invocations/provider-local replays of shared helpers.
- Test-only exports/globals/wrappers; production paths called only by tests.
- Mocks implementing asserted behavior or pretending distinct APIs match.
- Fixtures supplying receipts, admission, callback ordering owner should produce;
persistence asserted against store path never writes.
- Capability flags asserted instead of promised delivery/acknowledgement.
- Negative controls passing on unrelated guard or unreachable rejection.
- Names promising behavior inputs/assertions never exercise.
## Retention bar
- Keep independent public API, SDK, protocol, config, migration, storage,
security, platform, default, prompt-byte, generated cross-language, package,
release, architecture contracts. Keep observable ordering + credible regressions.
- Source inspection valid when cheapest independent guard of user-facing key,
byte, path, and survives identifier-only refactoring.
- Static/slow ≠ deletion reason. Baseline failure may be product bug:
reproduce + repair owner; never delete test for it.
## Candidate evidence
- Before editing: read full test, production owner, entry point, callers,
callees, siblings, overlapping proof, CI routing, history. Dependency-backed
claims: inspect dependency source/types. Read root/scoped instructions.
- Record each candidate's:
- Exact name/location + detectable failure.
- Non-test callers of covered production/support seam.
- Stronger remaining owner proof, or why none needed.
- History/reason it exists + deletion it unlocks.
- Risk + focused validation command.
- Missing evidence: no deletion. Discovery read-only; report before edits.
Prefer few high-confidence candidates. Parallel discovery may split independent
owners.
- Delete obsolete seams/dead paths; move retained regressions to owner;
consolidate generic package/dependency contracts.
- Prefer net-negative production LOC. Never add replacement implementation tests
or turn uncertainty into cleanup.
## Validation and handoff
- Follow the repository testing guide.
Never edit a checkout during an active test run. Run owner and sibling tests.
- Removed source greps/plan assertions: exercise owning executable/dry-run.
- Run targeted formatting, `git diff --check`, and required changed-file checks.
- Before landing removal batches, run the repository's unused dependency, file,
and export checks when available. Removed tests can orphan support code.
- Inspect `git diff --numstat`; review the full diff.
- Report causes, removed categories/seams, retained false positives, proof/cost,
production versus test/support LOC, PR state, follow-ups.
- Commit/push/land only when authorized through native maintainer workflow.
Refresh from main + rediscover before another audit batch.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |