Quality Gates (Parallel Execution)

After @builder completes, BOTH quality gates run simultaneously.

Parallel Execution Pattern

                    @builder completes
                           │
           ┌───────────────┴───────────────┐
           ▼                               ▼
    ┌─────────────┐                 ┌─────────────┐
    │ @validator  │                 │  @tester    │
    │ Code Quality│                 │ UX Quality  │
    ├─────────────┤                 ├─────────────┤
    │ ✓ TypeScript│                 │ ✓ E2E Tests │
    │ ✓ Unit Tests│                 │ ✓ Screenshots│
    │ ✓ Security  │                 │ ✓ A11y      │
    │ ✓ Consumers │                 │ ✓ Perf      │
    └──────┬──────┘                 └──────┬──────┘
           │                               │
           └───────────────┬───────────────┘
                      SYNC POINT
                           │
                   Apply Decision Matrix

Decision Matrix (MANDATORY)

Both agents MUST complete before applying this matrix:

@validator	@tester	NEXT ACTION
✅ APPROVED	✅ APPROVED	PROCEED to @scribe
✅ APPROVED	🔴 BLOCKED	RETURN to @builder (with @tester feedback)
🔴 BLOCKED	✅ APPROVED	RETURN to @builder (with @validator feedback)
🔴 BLOCKED	🔴 BLOCKED	RETURN to @builder (with MERGED feedback from both)

Rules:

• You MUST wait for BOTH agents before deciding
• @scribe can ONLY be called when BOTH gates are APPROVED
• If ANY gate is BLOCKED → back to @builder with specific feedback
• Maximum 3 retry cycles before escalation to user

@validator Checks (Code Quality)

Check	Tool	Blocking?
TypeScript compilation	npx tsc --noEmit	YES
Unit tests pass	npm test	YES
No security vulnerabilities	Static analysis	YES
Consumer impact verified	Type-check consumers	YES if API change
Code style / linting	npm run lint	NO (warning only)

Minimum output: 400 characters Required sections: Summary, Checks Performed, Issues Found, Verdict

@tester Checks (UX Quality)

Check	Tool	Blocking?
E2E tests pass	Playwright MCP	YES
Screenshots at 3 viewports	Playwright MCP	YES
Console errors captured	Browser console	YES (errors only)
WCAG 2.1 AA compliance	a11y checks	YES
Core Web Vitals	LCP, CLS, INP, FCP	NO (warning if poor)

Screenshot viewports:

• Mobile: 375px
• Tablet: 768px
• Desktop: 1920px

Minimum output: 800 characters Required sections: Summary, Screenshots Table, Console Errors, Performance Metrics, Verdict

Execution with Worktree Isolation

For conflict-free parallel execution, use isolation: worktree (native Claude Code subagent frontmatter — already set in validator.md/tester.md; no repo tooling required):

Task tool → subagent_type: "validator", isolation: "worktree"
Task tool → subagent_type: "tester", isolation: "worktree"

Both agents get their own git worktree — no file conflicts possible.

Performance

Mode	Duration	Improvement
Sequential	8–12 min	baseline
Parallel	5–7 min	40% faster
Parallel + Worktree	5–7 min	40% faster + zero conflicts

Fail-Safe (Graceful Degradation)

If an agent crashes (MCP failure, timeout):

1. Full Report — Normal operation, all checks pass
2. Partial Report — Some checks completed, others failed/timed out
3. Failure Report — Agent crashed, structured error output

Failure report includes: error type, suggested action (retry/escalate/skip), completion percentage.

Agent Return Contract

Each agent writes a full report to reports/vX.Y.Z/sprint-NN/<NN>-<agent>-report.md (canonical numbering: docs/templates/REPORT_TEMPLATES.md). The return message to the Orchestrator is the structured verdict only — separate from the on-disk report.

STATUS: APPROVED | BLOCKED | DONE
- finding 1 (one line max)
- finding 2
- finding 3
report: <absolute path to full report>

Important: scripts/validate-agent-output.js enforces min-length on the file, not the return message. This contract does not change validation behavior. Min-lengths remain: architect 1000, api-guardian 800, builder 500, validator 400, tester 800, scribe 300, github-manager 200.