quality-gate

name	quality-gate
description	Pre-push quality gate checklist. Use when: before any git push, before opening a PR, after implementing a full feature. Runs lint, type-check, tests, and a security scan as a single sequential gate. Fails fast — does not proceed past a failing stage.

Quality Gate Skill

When to Use

• Before every git push (mandatory)
• Before opening or merging a PR
• After implementing a complete feature
• When the user runs /quality-gate
• As the final step before any deployment

This gate answers one question: Is this code safe to ship?

---

Gate Sequence

Run stages in this exact order. Stop on first failure. Do not report "gate passed" if any stage fails.

Stage 1: Lint
Stage 2: Type-Check
Stage 3: Tests
Stage 4: Security Scan
Stage 5: Final Verdict

---

Stage 1: Lint

Catch style violations, unused imports, and obvious errors before they reach review.

Run

# JavaScript / TypeScript
npx eslint . --ext .ts,.tsx,.js,.jsx --max-warnings 0

# Python
ruff check .

# Go
golangci-lint run

# Ruby
rubocop

Adapt the command to the project's package.json scripts if a lint script exists (prefer npm run lint).

Pass Criteria

• Zero lint errors
• Zero lint warnings (warnings left in = technical debt shipped)

On Failure

Fix all lint issues. Do not disable lint rules without explicit user approval. Do not proceed to Stage 2 until Stage 1 is clean.

---

Stage 2: Type-Check

Catch type errors that slip past lint.

Run

# TypeScript
npx tsc --noEmit

# Python (mypy)
mypy .

# Go (built into build)
go build ./...

Adapt to the project's typecheck script if one exists in package.json.

Pass Criteria

• Zero type errors
• --noEmit used (do not emit output files as a side effect)

On Failure

Fix all type errors. Do not use // @ts-ignore or any to silence errors without user approval. Do not proceed to Stage 3 until Stage 2 is clean.

---

Stage 3: Tests

Validate correctness end-to-end.

Run

# Jest / Vitest
npm test -- --run --reporter=verbose

# pytest
pytest -v --tb=short

# Go
go test ./... -v

# RSpec
bundle exec rspec --format documentation

Use the project's test script if defined (prefer npm test, yarn test, etc.).

Pass Criteria

• All tests pass (0 failures, 0 errors)
• No skipped tests that should be running (review test.skip / xit / pending)
• If coverage is configured: coverage does not decrease below baseline

On Failure

Run the failing tests in isolation to diagnose. Fix the failure. Re-run the full suite. Do not proceed to Stage 4 until Stage 3 is clean.

---

Stage 4: Security Scan

Quick surface-level scan for critical issues introduced in this changeset.

4A: Dependency Audit

# Node.js
npm audit --audit-level=high

# Python
pip-audit

# Ruby
bundle audit check --update

4B: Secrets Scan

Check that no credentials, tokens, or keys were accidentally committed:

# If git-secrets or gitleaks is available
gitleaks detect --source . --no-git

# Otherwise: manual grep
grep -rn "password\s*=\s*['\"]" --include="*.ts" --include="*.js" --include="*.py" .
grep -rn "api_key\s*=\s*['\"]" --include="*.ts" --include="*.js" --include="*.py" .
grep -rn "secret\s*=\s*['\"]" --include="*.ts" --include="*.js" --include="*.py" .

4B2: Dependency Supply Chain Audit (HIGH PRIORITY)

If package.json / requirements.txt / Gemfile / go.mod changed, invoke @security agent BEFORE pushing.

Do NOT assume the quality gate audit is sufficient. Security performs:

• Typosquatting detection (name similarity to popular packages)
• Maintainer verification (unexpected changes, abandoned packages)
• Transitive dependency audit (indirect deps)
• SBOM generation (artifacts for compliance/inventory)
• Lock file integrity validation

See Security agent documentation for full dependency review process.

4C: Quick OWASP Check (for changed files only)

For each file modified in this push, verify:

• User input is validated before use
• No raw SQL string concatenation
• No eval() or equivalent dynamic code execution
• No hardcoded credentials
• Auth checks present on any new route/endpoint

For a full audit, invoke the @security agent separately.

Pass Criteria

• Zero HIGH or CRITICAL dependency vulnerabilities (direct)
• Transitive dependencies also checked (npm audit includes this)
• No secrets found in source files
• No obvious OWASP violations in changed files
• If dependencies changed: SBOM generated + @security agent cleared before push

On Failure

• Dependency vulnerability: Update the package or document the exception with user approval. Do not push with unmitigated HIGH/CRITICAL CVEs.
• Secret found: Remove immediately, rotate the credential, force-push if already committed.
• OWASP violation: Fix before pushing. Tag @security for a full audit if the scope is large.

---

Stage 5: Final Verdict

Only after all four stages pass:

## Quality Gate: PASSED ✅

| Stage        | Result  | Notes                        |
|:-------------|:--------|:-----------------------------|
| Lint         | ✅ Pass | 0 errors, 0 warnings         |
| Type-Check   | ✅ Pass | 0 errors                     |
| Tests        | ✅ Pass | 142 passed, 0 failed         |
| Security     | ✅ Pass | 0 vulns, 0 secrets           |

**Safe to push.**

If any stage fails, output:

## Quality Gate: BLOCKED ❌

| Stage      | Result  | Notes                              |
|:-----------|:--------|:-----------------------------------|
| Lint       | ✅ Pass |                                    |
| Type-Check | ❌ FAIL | 3 type errors in auth/session.ts   |
| Tests      | ⏭ Skip | Blocked by Stage 2 failure         |
| Security   | ⏭ Skip | Blocked by Stage 2 failure         |

**Do NOT push. Fix Stage 2 failures first.**

---

Integration with Security Agent

The quality gate's Stage 4 is a quick surface scan, not a full audit.

Mandatory @security review for:

• Any dependency changes (stages 4A2 covers the full supply chain audit)
• Auth, payments, file uploads
• New API surface
• Any HIGH/CRITICAL findings in 4A or 4B

Always invoke @security after the gate if any of the above apply.

The gate catches the obvious. The Security agent catches the subtle. Supply chain attacks are sophisticated — never skip the Security agent review for dependency changes.