spec-verify

name	spec-verify
description	Verify a completed implementation against its spec. Checks all acceptance criteria, runs tests, validates coverage. Use after spec-implement is done.

Spec Verify Workflow

Verify that a completed implementation satisfies its spec.

Steps

1. Read the spec - load requirements.md and design.md
2. Run verification checks against each requirement
3. Generate a verification report
4. Flag any gaps for remediation

Verification Checklist

For each requirement in requirements.md:

Functional Verification

• Every acceptance criterion has a passing test
• Edge cases identified in design.md are handled
• Error paths are tested (not just happy path)

Non-Functional Verification

• Performance requirements met (if specified)
• Security requirements met (no secrets, proper auth, input validation)
• Accessibility requirements met (if frontend)

Code Quality

• Tests pass (npm test / pytest / equivalent)
• Type checking passes (if applicable)
• No lint errors
• No hardcoded values that should be configurable

Documentation

• Code comments explain WHY, not WHAT
• API changes documented
• ADR written if architectural decisions were made
• Changelog entry added

Verification Report

After checking, append a verification section to the spec folder:

<!-- Append to tasks.md -->

## Verification Results

**Date:** YYYY-MM-DD
**Status:** ✅ PASSED | ⚠️ PARTIAL | ❌ FAILED

### Requirements Coverage
| Requirement | Status | Notes |
|-------------|--------|-------|
| REQ-1 | ✅ | All criteria met |
| REQ-2 | ⚠️ | Missing edge case test |

### Gaps Found
- Gap 1: ...
- Gap 2: ...

### Recommendation
Ready for archive / Needs remediation on: ...

Rules

1. Be strict - a requirement is only "met" if ALL its acceptance criteria pass
2. Check tests exist - don't just trust that code works; verify tests cover it
3. Flag scope creep - if implementation added things not in the spec, note it
4. Recommend next step - either spec-archive (if passed) or remediation tasks