Claude-skill-registry implementation-verification

Use after implementation complete to verify all tasks done, update roadmap, run full test suite, and create final report - ensures implementation completeness before finishing development branch

install
source · Clone the upstream repo
git clone https://github.com/majiayu000/claude-skill-registry
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/majiayu000/claude-skill-registry "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/data/implementation-verification" ~/.claude/skills/majiayu000-claude-skill-registry-implementation-verification && rm -rf "$T"
manifest: skills/data/implementation-verification/SKILL.md
source content

Implementation Verification

What It Does

  1. Verifies all tasks marked complete
  2. Updates product roadmap
  3. Runs entire test suite
  4. Creates final verification report
  5. Reports pass/fail with specific details

Pass = ready to finish. Fail = fix issues first.

The Process

Step 1: Verify All Tasks

SPEC="[provided by workflow]"

cat "$SPEC/tasks.md"

# Count tasks
TOTAL=$(grep -c "^- \[.\]" "$SPEC/tasks.md")
COMPLETE=$(grep -c "^- \[x\]" "$SPEC/tasks.md")
INCOMPLETE=$(grep "^- \[ \]" "$SPEC/tasks.md")

For each incomplete task:

TASK_ID=$(echo "$INCOMPLETE" | grep -o "[0-9]\.[0-9]")

# Check for implementation report
ls "$SPEC/implementation/" | grep -i "$TASK_ID"

# Spot check code
# [Search relevant files]

If appears complete but not marked:

Task [N.M]: [Description]

Status: ⚠️ Appears complete, not marked

Evidence:
- Implementation report: [Yes/No]
- Code changes: [Yes/No]
- Files: [list]

Marking complete...

Update:

sed -i "s/^- \[ \] $TASK_ID/- [x] $TASK_ID/" "$SPEC/tasks.md"

If truly incomplete:

Task [N.M]: [Description]

Status: ❌ NOT COMPLETE

Evidence:
- No report
- No code changes
- Missing: [list]

Noted in verification report.

Step 2: Update Roadmap

cat specs/product/roadmap.md

SPEC_NAME=$(basename "$SPEC" | sed 's/^[0-9-]*//')

# Find matching item
grep -i "$SPEC_NAME" specs/product/roadmap.md

If found:

LINE=$(grep -n "^[0-9]*\. \[ \].*$SPEC_NAME" specs/product/roadmap.md | cut -d: -f1)

if [ ! -z "$LINE" ]; then
  sed -i "${LINE}s/\[ \]/\[x\]/" specs/product/roadmap.md
fi

Announce:

✅ Roadmap updated

Marked complete: [Item]

If not found:

ℹ️  No matching roadmap item

This might be:
- Custom feature (not from roadmap)
- Already marked complete
- Named differently

Skipping roadmap update.

Step 3: Run Full Test Suite

Detect test command:

if [ -f "package.json" ]; then
  TEST_CMD="npm test"
elif [ -f "Cargo.toml" ]; then
  TEST_CMD="cargo test"
elif [ -f "requirements.txt" ] || [ -f "pyproject.toml" ]; then
  TEST_CMD="pytest"
elif [ -f "go.mod" ]; then
  TEST_CMD="go test ./..."
else
  # Check CLAUDE.md or tech-stack.md
  TEST_CMD="[detect from docs]"
fi

# Run tests
$TEST_CMD 2>&1 | tee test-output.log
TEST_EXIT=$?

Parse results:

PASSING=$(grep -c "PASS\|✓\|ok" test-output.log)
FAILING=$(grep -c "FAIL\|✗\|not ok" test-output.log)
FAILED_TESTS=$(grep -B 2 "FAIL\|✗" test-output.log)

Step 4: Create Final Report

cat > "$SPEC/verification/final-verification.md" <<EOF
# Verification Report: [Spec Name]

**Spec:** \`[spec-name]\`
**Date:** $(date +%Y-%m-%d)
**Verifier:** implementation-verification
**Status:** [✅ Passed / ⚠️ Issues / ❌ Failed]

---

## Executive Summary

[2-3 sentence overview]

---

## 1. Tasks Verification

**Status:** [✅ All Complete / ⚠️ Issues]

### Completed Tasks

- [x] Task Group 1: [Title]
  - [x] Subtask 1.1
  - [x] Subtask 1.2

[List all completed]

### Incomplete or Issues

[If any:]
- [ ] Task [N.M]: [Description]
  - Issue: [What's missing]
  - Evidence: [Checked]

[If all complete:]
None - all verified complete.

---

## 2. Documentation

**Status:** [✅ Complete / ⚠️ Issues]

### Implementation Reports

- [x] Group 1: \`implementation/1-[name].md\`
- [x] Group 2: \`implementation/2-[name].md\`

### Missing

[If any:]
- [ ] Group [N] report

[If all present:]
None - all reports present.

---

## 3. Roadmap Updates

**Status:** [✅ Updated / ℹ️ No Match / ⚠️ Issues]

### Updated Items

[If updated:]
- [x] [Roadmap item marked complete]

[If no match:]
No matching item - may be custom feature.

---

## 4. Test Suite Results

**Status:** [✅ Passing / ⚠️ Failures / ❌ Critical]

### Summary
- **Total:** [$TOTAL_TESTS]
- **Passing:** [$PASSING]
- **Failing:** [$FAILING]

### Failed Tests

[If failures:]
1. [Test name]
   - File: [file]
   - Error: [snippet]

[If passing:]
None - all tests passing ✅

---

## 5. Quality Assessment

**Status:** [✅ Good / ⚠️ Concerns / ❌ Issues]

### Strengths
- [Positive aspect 1]
- [Positive aspect 2]

### Concerns
[If any:]
- ⚠️ [Concern]

---

## Conclusion

**Assessment:** [Detailed assessment]

[If passed:]
Implementation complete and verified. All tasks done,
tests passing, roadmap updated. Ready for completion.

[If issues:]
Implementation has [X] incomplete tasks and [Y] failing
tests that must be resolved.

**Recommendation:** [Next steps]
EOF

rm -f test-output.log

Step 5: Present Results

If PASSED:

✅ Implementation Verification PASSED

Summary:
- Tasks: [X/X] complete (100%)
- Roadmap: Updated ✅
- Tests: [Y] passing, 0 failures
- Documentation: Complete ✅

Report: verification/final-verification.md

🎉 Implementation complete!

Ready to finish development work.

If ISSUES:

⚠️  Verification PASSED (with issues)

Summary:
- Tasks: [X/Y] complete ([Z] marked manually)
- Roadmap: Updated ✅
- Tests: [A] passing, 0 failures
- Documentation: [B] reports ([C] missing)

Minor issues:
1. [Issue]

Report: verification/final-verification.md

Continue to finish?

If FAILED:

❌ Implementation Verification FAILED

Summary:
- Tasks: [X/Y] complete ([Z] incomplete)
- Roadmap: [Status]
- Tests: [A] passing, [B] FAILING
- Documentation: [C] reports ([D] missing)

Critical:
1. [X] incomplete tasks:
   - Task [N.M]: [Description]

2. [B] failing tests:
   - [Test 1]
   - [Test 2]

Report: verification/final-verification.md

Options:
1. Review full report
2. Fix failing tests
3. Complete remaining tasks
4. Return to implementation

What next?

WAIT if failed.

Handling Failures

If tests fail:

[B] tests failing.

Failed categories:
- [X] in [component]
- [Y] in [component]

These are:
[Related to incomplete / Regressions / New issues]

Suggested:
[Specific recommendation]

Proceed?

If tasks incomplete:

[Z] tasks not complete:

1. Task [N.M]: [Description]
   - Missing: [What's needed]

Options:
1. Complete these now
2. Mark deferred (document why)
3. Review/adjust tasks

What next?

Red Flags

Never:

  • Mark tasks complete without evidence
  • Skip full test suite
  • Ignore test failures
  • Skip roadmap update
  • Assume documentation exists

Always:

  • Verify each task with evidence
  • Run complete test suite
  • Parse test results accurately
  • Update roadmap if match found
  • Document all findings
  • Present clear next steps

Integration

Called by:

  • spec-implementation-workflow
    (Step 6)

Returns to:

  • spec-implementation-workflow
    with status

Creates:

  • [spec]/verification/final-verification.md

Updates:

  • [spec]/tasks.md
    - Marks verified-but-unmarked tasks
  • specs/product/roadmap.md
    - Checks off items

Next steps:

  • If passed:
    finishing-development-branch
  • If failed: Return to implementation or present options