Where this fits
You've just finished a feature branch, the CI is green, and you're about to merge to main. But a nagging question remains: did you leave untested code paths, duplicate logic, or inconsistent error handling? The gza-code-review-full skill automates this pre-release check, giving you a structured, reproducible quality report without manual flailing.
Why agents benefit
* Runs 10 structured checks in one pass: test coverage, duplication, component interactions, error handling, API consistency, config audit, logging, resource management, and type safety.
* Writes findings to a file (reviews/) for traceability and comparison across releases.
* Uses only allowed tools (Read, Glob, Grep, Bash with sandboxed run commands), ensuring safe execution in any agent environment.
* Produces actionable, not generic, output — each check maps to concrete code patterns, not vague "looks good."
* Works with any model — the skill is model-agnostic and names the reviewing model in the output file for auditability.
Practical scenarios
Pre-release health check
Before tagging a release, invoke the skill to scan the entire codebase. It reports unit and functional test coverage gaps, duplicate functions, and risky component coupling — exactly what you'd want a senior engineer to flag.
Onboarding a new team member
New developer on the team? Run this skill against a PR branch to give them a concrete, non-judgmental list of code quality improvements before review.
Technical debt review sprint
Dedicate a sprint to quality. Run the skill on each module, compare timestamped reports, and track reduction in duplication and uncovered code over time.
Add it to your agent workflow
Install and run in one step:
npx skills add mhawthorne/gza --skill gza-code-review-full
Example invocation in your agent's configuration:
- skill: gza-code-review-full
config:
output_dir: ./reviews
model_identifier: claude-sonnet-4
For full details, see the skill page on OpenAgentSkill.
Compare before adopting
Evaluate quality: does the skill detect meaningful patterns (e.g., missing error handling, tight coupling) or just surface-level issues? Check maintenance freshness: last commit date, stars (11), and responsiveness to issues. Compare with alternatives like code-review or code-quality skills — does gza-code-review-full cover your specific pre-release workflow? Also assess how it integrates with your CI or agent runner.
Why it is worth tracking
With 11 GitHub stars and a clear 10-point review structure, this skill serves a focused need: reproducible pre-release quality assessments. Monitor its star growth and release cadence. If your team values structured code audits before merge, this skill deserves a trial run.