From specialist-agent
Runs npm test, lint, build, and type check commands to verify code before commits, PRs, or task completion claims. Requires fresh output evidence.
npx claudepluginhub herbertjulio/specialist-agent --plugin specialist-agentThis skill is limited to using the following tools:
Evidence before claims. Always.
Enforces running verification commands like tests, build, lint, type-check and confirming output before claiming task completion, fixes, or commit/PR readiness.
Enforces five-step verification (identify/run/read/verify/claim) of commands like npm test/build/lint before claiming code complete, fixed, or passing. Builds trust via evidence.
Enforces running fresh verification commands like tests, builds, or linters before claiming work is complete, fixed, passing, committing, or creating PRs.
Share bugs, ideas, or general feedback.
Evidence before claims. Always.
Target: $ARGUMENTS
NO COMPLETION CLAIMS WITHOUT FRESH VERIFICATION EVIDENCE
If you haven't run the verification command in THIS message, you cannot claim it passes.
Violating the letter of this rule is violating the spirit of this rule.
BEFORE claiming any status:
1. IDENTIFY - What command proves this claim?
2. RUN - Execute the FULL command (fresh, complete, not cached)
3. READ - Full output, check exit code, count failures
4. VERIFY - Does output ACTUALLY confirm the claim?
→ If NO: State actual status with evidence
→ If YES: State claim WITH the evidence
5. ONLY THEN - Make the claim
Skip any step = lying, not verifying.
| Claim | Required Command | NOT Sufficient |
|---|---|---|
| "Tests pass" | npm test output: 0 failures | Previous run, "should pass" |
| "Lint clean" | npm run lint output: 0 errors | "I fixed the warning" |
| "Build succeeds" | npm run build exit code 0 | "Lint passed so build should too" |
| "Bug fixed" | Test that reproduces bug: PASS | "Code changed, should work" |
| "Types correct" | npx tsc --noEmit exit code 0 | "I added the types" |
| "No regressions" | Full test suite: 0 failures | "Only changed one file" |
| "Requirements met" | Line-by-line checklist verified | "Tests pass" |
npm test 2>&1
# Read FULL output
# Count: X passed, Y failed, Z skipped
# Exit code: 0 or non-zero
npm run build 2>&1
# Check: exit code 0?
# Check: no error output?
npm run lint 2>&1
# Count: 0 errors, 0 warnings
npx tsc --noEmit 2>&1
# Count: 0 errors
Run ALL of the above in sequence. Stop at first failure.
Auto-detect what to verify based on package.json scripts.
──── /verify ────
Command: npm test
Exit code: 0
Result: 42/42 tests passed, 0 failed
VERIFIED ✓ - All tests pass (evidence above)
Or:
──── /verify ────
Command: npm test
Exit code: 1
Result: 40/42 tests passed, 2 failed
FAILED ✗ - 2 tests failing:
- src/auth.test.ts: "should validate token" - Expected true, got false
- src/user.test.ts: "should create user" - Timeout after 5000ms
DO NOT claim completion. Fix failures first.
If you catch yourself thinking:
ALL of these mean: RUN THE COMMAND. READ THE OUTPUT.
| Excuse | Reality |
|---|---|
| "Should work now" | RUN the verification |
| "I'm confident" | Confidence is not evidence |
| "Just this once" | No exceptions, ever |
| "Lint passed" | Lint is not build is not test |
| "Agent said success" | Verify independently |
| "I only changed a comment" | Comments can break builds (JSX, directives) |
| "Previous run passed" | State changed since then. Run again. |
| "Partial check is enough" | Partial proves nothing |
| "Too slow to run full suite" | Slow verification beats fast regression |
This skill is a cross-cutting concern. It applies to: