What is the primary purpose of the review-implementation skill?

It acts as a final quality gate to verify that the code written matches the technical specifications and success criteria defined in the project epic before a task is marked as finished.

Can I use this skill while I am still writing code?

No, this skill is designed to be used after all tasks in a plan are completed but before claiming the work is finished or merging the branch.

How does this differ from a standard code review?

It is a highly structured, automated self-review process that uses a 'Google Fellow' level of scrutiny to check for specific anti-patterns, stubs, and edge cases that automated tests might miss.

Does it only check if the tests pass?

No, it goes beyond testing by verifying documentation, safety patterns, error handling, and whether every specific requirement in the task description was actually implemented.

What tools does this skill require?

It leverages the hyperpowers task tracking system (bd), ripgrep (rg) for code searching, and a test-runner agent to execute your project's specific build and test commands.

Implementation Reviewer

Name: Implementation Reviewer
Author: withzombies

bywithzombies

•

Security & Testing

Verifies completed implementations against project specifications using high-rigidity quality gates and automated code audits.

The review-implementation skill serves as a rigorous quality assurance layer for Claude Code, ensuring that every task within a development epic is fully completed according to the original specification. By adopting the persona of a veteran SRE, it enforces a standardized four-step verification process that includes reading full source files, running automated linting and test suites via specialized agents, and validating success criteria with concrete evidence. It is specifically designed to catch implementation gaps, anti-patterns, and unfinished 'todo' stubs before code is promoted to production or a development branch is closed.

Key Features

01Full source code audit that analyzes entire files rather than just git diffs

02Automated detection of TODOs, stubs, and unsafe code patterns using ripgrep

03Deep integration with 'bd' task tracking to verify individual success criteria

04Multi-agent quality gates for running test suites, linters, and pre-commit hooks

0513 GitHub stars

06Evidence-based reporting that provides clear approval or detailed gap analysis

Use Cases

01Auditing code for production-readiness to ensure compliance with SRE-level safety standards

02Ensuring complex multi-task epics are 100% compliant with the initial project specification

03Final verification of a feature branch before triggering a merge or completion workflow

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add withzombies/hyperpowers review-implementation

For use in Claude.ai and ChatGPT

Download Skill