skills.rest
lossyrob
lossyrobCOMMUNITY
Explore all skills by this author →
𝕏@lossyrobGitHub

paw-final-review

Compare implementation diffs against project specifications to identify gaps.

39|6|Updated Oct 8, 2025
Search Tags:#multi-model#review#diff#spec#PR#code-review#PAW
One-click install
npx skills add https://github.com/lossyrob/phased-agent-workflow --skill paw-final-review
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: paw-final-review
Source: https://github.com/lossyrob/phased-agent-workflow/tree/main/skills/paw-final-review
Command: npx skills add https://github.com/lossyrob/phased-agent-workflow --skill paw-final-review

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

The Final Agent Review skill provides a structured, automated check that compares the complete implementation diff against project specifications and plans, catching gaps before a Final PR is created.

Core Features & Use Cases

  • Review implementation against spec for correctness, patterns, and issues
  • Multi-model parallel review with synthesis (CLI only)
  • Single-model review (CLI and VS Code)
  • Interactive or auto-apply resolution modes
  • Generate review artifacts in .paw/work/<work-id>/reviews/
  • Works after all implementation phases are complete, before Final PR creation
  • Integrates with PAW sessions to preserve user interactivity during decisions

Quick Start

Use the paw-final-review skill within a PAW session to analyze the current implementation diff against the spec and produce a REVIEW.md artifact for the PR flow.

Frequently Asked Questions about paw-final-review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate code review against a spec before creating a PR?▼

Automating pre-PR code review against a spec involves comparing the full implementation diff against project specifications to identify gaps. You can use a PAW final review session to read Spec.md and ImplementationPlan.md, analyze the diff, and export a REVIEW.md artifact.

Can I use multi-model code review with synthesis in VS Code?▼

Multi-model parallel review with synthesis is only supported in the CLI. VS Code supports single-model review only, while both CLI and VS Code offer interactive or auto-apply resolution modes during the PAW review workflow.

What files are needed to compare implementation diff for spec compliance?▼

Comparing implementation diff for spec compliance requires Spec.md, ImplementationPlan.md, and CodeResearch.md. The review process reads these project files, gathers the complete diff, and evaluates correctness, patterns, and issues before PR creation.

How do I generate review artifacts for a PAW workflow session?▼

Generating review artifacts in a PAW workflow session happens automatically when the final review analyzes the diff against the spec. It exports artifacts to a designated .paw/work directory path, producing a REVIEW.md file to support the PR flow.

When should I run a pre-PR spec compliance check in my workflow?▼

A pre-PR spec compliance check runs during the Final Review phase after all implementation phases are complete and before Final PR creation. It compares the complete implementation diff against project specifications and plans to catch gaps.

Does the PAW final review skill require external dependencies?▼

The PAW final review skill operates with no external dependencies. It works within PAW sessions to preserve user interactivity during decisions and uses existing project files like Spec.md to execute the diff comparison.

Related Skills

Explore More Software Engineering Skills→
lossyroblossyrob
39

paw-planning

Generate a phased ImplementationPlan.md with milestones and acceptance criteria.

Community
Basic
lossyroblossyrob
39

paw-review-workflow

Orchestrate PAW activity skills to generate structured PR review feedback.

Community
Advanced
CUexterCUexter
2

blackbox

Orchestrate multi-model coding tasks via the Blackbox CLI.

Community
Advanced
LFDT-NightstreamLFDT-Nightstream
29

multi-ai-council

Coordinate multi-provider AI passes for hard technical questions.

Official
Intermediate