Agent Skills: Review Implementation Against Spec

Run comprehensive code review before commits, PRs, or releases. Uses code-reviewer agent with tiered security, performance, and quality checks.

UncategorizedID: melodic-software/claude-code-plugins/review

Install this agent skill to your local

pnpm dlx add-skill https://github.com/melodic-software/claude-code-plugins/tree/HEAD/plugins/tac/skills/review

Skill Files

Browse the full folder contents for review.

Download Skill

Loading file tree…

plugins/tac/skills/review/SKILL.md

Skill Metadata

Name
review
Description
Compare implementation against specification to verify alignment. Use after implementation to verify requirements are met.

Review Implementation Against Spec

Compare the implementation against the specification to verify alignment.

Variables

  • spec_file: $1 - Path to the specification file

Purpose

Review answers: "Is what we built what we asked for?"

This is different from testing. We know functionality works. We're validating alignment with the original specification.

Instructions

1. Read the Specification

Read the spec file to understand:

  • Original requirements
  • Success criteria
  • Expected behavior
  • Acceptance criteria

2. Analyze Changes

Compare implementation against spec:

# See all changes
git diff origin/main

# See changed files
git diff origin/main --stat

# See specific file changes
git diff origin/main -- path/to/file

3. Capture Screenshots (1-5)

Take screenshots of critical functionality paths:

  1. Initial State: Before interaction
  2. Key Actions: After significant user actions
  3. Final State: End result

Name format: 01_descriptive_name.png, 02_descriptive_name.png

4. Compare Against Spec

For each requirement in the spec:

  • Is it implemented?
  • Does it match the specification?
  • Are there any deviations?

5. Classify Issues

For each issue found, classify by severity:

| Severity | Description | Action | | --- | --- | --- | | blocker | Prevents release, harms UX | Must fix | | tech_debt | Quality issue, feature works | Document | | skippable | Polish, preference | Note only |

Output Format

Return ONLY JSON:

{
  "success": true,
  "review_summary": "2-4 sentence summary of review findings",
  "review_issues": [
    {
      "issue_description": "Clear description of the issue",
      "issue_resolution": "How to fix it",
      "issue_severity": "blocker"
    },
    {
      "issue_description": "Another issue",
      "issue_resolution": "Resolution approach",
      "issue_severity": "tech_debt"
    }
  ],
  "screenshots": [
    "path/to/01_initial_state.png",
    "path/to/02_after_action.png"
  ]
}

Success Criteria

  • success: true when NO blocker issues exist
  • success: false when ANY blocker issues exist

Review can succeed with tech_debt and skippable issues.

Classification Guidelines

Blocker if:

  • Feature doesn't work as specified
  • UI doesn't match requirements
  • Critical functionality missing
  • User experience is harmed

Tech Debt if:

  • Feature works but implementation is suboptimal
  • Code quality issues
  • Missing optimization
  • Console warnings

Skippable if:

  • Minor visual differences
  • Polish improvements
  • Subjective preferences
  • Nice-to-have enhancements

Integration with Workflow

This command is part of the SDLC:

/plan → /implement → /test → /review → /patch (if needed) → /document

If blockers are found, use /patch to create targeted fixes.