SKILL.md
Validate & Auto-Fix
Run the project validation pipeline, auto-fix trivial issues, and re-run until green or a real failure is found.
Rules
- CRITICAL: Never use
go testdirectly - ALWAYS usemake test(handles envtest, CRDs, build flags) - Never modify production logic to fix a test. Only fix test expectations, imports, formatting.
- Never skip or delete a failing test.
- Stop after 3 auto-fix cycles to avoid loops.
- Report real failures clearly; do not attempt speculative fixes.
Step 1: Run Unit Tests
make test 2>&1 | tail -60Important: The Makefile handles essential setup (envtest, CRDs, build flags) that go test doesn't.
If all pass, proceed to Step 3 (linting). If failures occur, classify each failure (see Step 2).
Step 2: Classify and Fix Failures
For each failure, determine its type:
Auto-fixable (fix immediately, then re-run Step 1):
| Type | Fix |
|---|---|
| Test expects old constant value | Update assertion to match new value |
| Test uses renamed function | Update function name in test |
| Import error from refactor | Update import path |
| Missing cleanup in test | Add cleanup or use existing cleanup helpers |
Real failures (do not auto-fix):
- Logic errors in production code
- Assertion failures reflecting actual behavior regressions
- Reconciliation loop failures
- Context cancellation issues
- Failures in code you did not modify
For real failures: report the test name, file, error message, and stop.
Step 3: Run Linting
make lint 2>&1 | tail -40This runs: golangci-lint, go fmt, go vet, and custom checks.
If failures occur, apply the same classify-and-fix logic from Step 2. Common fixes at this stage:
| Type | Fix |
|---|---|
gofmt formatting |
go fmt ./... |
| Unused import | Remove the import |
| Unused variable | Remove the variable or use _ = variable if intentional |
| Missing error check | Add if err != nil { return err } |
| Ineffectual assignment | Remove or fix the assignment |
Re-run make lint after each fix. Proceed to Step 4 when green.
Step 4: Run E2E Tests (Optional)
Only if the user explicitly asks or if changes affect reconciliation logic:
make test-e2e 2>&1 | tail -30Requirements: Requires a running OpenShift/Kubernetes cluster with operator deployed.
E2E test failures are almost always real failures. Report and stop.
Step 5: Check Bundle and Manifests
If API changes were made (api/v1alpha1/), regenerate manifests:
make generate
make manifests
git diffIf there are differences, commit them:
git add api/ config/
git commit -m "Regenerate manifests"Step 6: Report
Report exactly:
make test: X passed / Y failed
- List any failing tests with brief error summary
make lint: pass/fail
- List any remaining lint errors
make test-e2e: (if run) X passed / Y failed- Auto-fixes applied:
- File: what was fixed (e.g., "internal/controller/utils/utils_test.go: updated assertion")
- Cycles used: N/3
If all green:
✅ All validation passed:
- make test: all tests passing
- make lint: no issues
Ready to commit or push.Do not include unrelated diagnostics or suggestions.