tt-a1i/archify / .agents/skills/test-value
test-value
VerifiedImprove test value and feedback cost when auditing slow, flaky, redundant or brittle tests; choosing tests during implementation; or reviewing test changes before a commit or PR. Keep ordinary implementation and pre-commit use scoped to the affected behavior, not a repository-wide audit.
| File | Kind | Size |
|---|---|---|
| LICENSE | other | 1.0 KB |
| SKILL.md | skill md | 9.4 KB |
| agents/openai.yaml | config | 320 B |
| references/audit.md | markdown | 3.6 KB |
| references/decision-examples.md | markdown | 5.0 KB |
references/audit.md
View on GitHub# Auditing Existing Tests
## Map the cost
Confirm the workspace, comparison baseline and uncommitted changes; read test entrypoints and CI configuration. Reuse current logs before collecting measurements that are genuinely missing.
For broad work, group tests by responsibility: core product behavior, shared algorithms, entrypoint integration, browser/platform behavior, filesystem and process safety, distribution/installation, generated artifacts and CI selection. Identify uninspected areas; one successful cut does not establish that the whole project is optimal.
Cost includes execution time, false-alarm investigation, fixture maintenance, blocked refactors, flaky reruns and resource contention. A lack of historical failures does not establish low value: rare data corruption can justify expensive protection.
Choose candidates from measured bottlenecks, frequent false alarms or current development friction. Avoid predetermined deletion ratios, coverage targets or universal time limits. An expensive test outside the daily loop may cost less than hundreds of small duplicates in every feedback cycle.
## Prove the boundary of a cut
For substantial reductions, keep a short, reviewable record:
- **Burden:** measured execution or maintenance cost, or a concrete false alarm.
- **Existing protection:** the actual mistake, entrypoints and implementation paths.
- **Change:** deletion, consolidation, rewritten assertion, fixture reuse or changed execution timing.
- **Surviving evidence:** how remaining tests detect the mistake; combinations intentionally no longer exercised.
- **Verification:** commands and outcomes, comparable timing, baseline failures and untested environments.
Prefer low-cost changes to a new test framework, dynamic dependency graph or layered runner. Selection machinery also creates maintenance obligations; introduce it only when existing tools cannot meet an evidenced need.
## Select execution by risk
Use the repository's risk distribution, not another project's command names or runtime versions.
- **Daily feedback:** core public behavior and affected regressions.
- **Expanded checks:** relevant integration and consumers of shared changes.
- **Platform and delivery checks:** distinct risks in real browsers, operating systems, installation, packaging and release identity.
- **Complete verification:** infrastructure changes, uncertain impact or the project's release requirements.
Before reducing platform matrices, distinguish runtime-compatibility smoke from platform-specific faults. Preserve actual evidence for platform differences. A mock or runner command list demonstrates dispatch, not native path behavior.
When changing affected-test selection, check new files, renames, deletions, shared dependencies and unknown paths. Classification failures must not silently become successful skips. Respect required checks; a request for faster feedback does not authorize changing branch protection or release policy.
## Work in batches
Complete a change and verification for one responsibility boundary before moving to the next. Stabilize unreliable measurements before claiming speedups. Track live process handles; an observation timeout does not justify starting another full run.
Reuse verification unaffected by later changes. When only proven redundant assertions are removed, run surviving behavior checks or reuse applicable results. Reducing test cost should not produce an ever-growing verification ritual.
At the end, identify retained protections, intentionally removed combinations and deferred candidates. Use an existing evidence location when useful rather than automatically creating a new body of governance documentation.
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: 463k | 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: 463k | 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: 463k | 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: 192k | Updated | Claude CodeCodex | Verified | Install |