Skip to content

njdaniel/conch / .claude/skills/test-and-fix

test-and-fix

Verified

Run the test suite, categorize every failure (regression / flake / environment), fix regressions, and never silently skip. Use whenever tests fail or before opening a PR.

test-and-fix

Procedure

  1. Run go test ./... (then make check for the full gate).
  2. Categorize every failure:
    • Regression — the code under test is wrong, or the change broke a real contract. Fix the code (or, if the test encoded a wrong expectation, fix the test and say so in the PR).
    • Flake — passes on rerun, timing/ordering dependent. Rerun with -count=5 -race to confirm. Don't just rerun-until-green: file an issue for the flake with the failure output, and fix it if the fix is small (deterministic clock, channel sync, t.TempDir).
    • Environment — missing tool, sandbox limits, network. Note it in the PR; make the test skip explicitly and loudly (t.Skipf with reason) only if the dependency is genuinely optional.
  3. Fix regressions in the current branch if in scope for the issue; otherwise file a blocking issue and stop.
  4. Re-run the full suite after fixes.

Never

  • Delete or t.Skip a failing test to get green without flagging it prominently in the PR description (CLAUDE.md rule 6).
  • Mark a regression as a flake because it passed once on rerun.
  • Ship with the approval-path E2E test failing — that is always a blocker (rule 3).
finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the workStars: 319kUpdated Claude CodeCodexVerifiedInstall
systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixesStars: 319kUpdated Claude CodeCodexVerifiedInstall
test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation codeStars: 319kUpdated Claude CodeCodexVerifiedInstall
skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.Stars: 196kUpdated Claude CodeCodexVerifiedInstall