Pre-Merge Code Quality Review
Review every change across correctness, readability, architecture, security, and performance before it reaches the main branch.
The skill explicitly covers input validation, secret handling, authorization, untrusted external data, dependency vulnerabilities, and lockfile review, and asks before deleting dead code. It does not systematically define least privilege, data-flow disclosure, user confirmation, rollback, or source attribution, so points are deducted.
The five-axis process, severity taxonomy, abnormal-path checks, and verification checklist are reasonably coherent. However, there are no dedicated tests, execution commands, or concrete diagnostic behavior for this skill's key paths; static evidence is limited and the score remains below the static cap.
The audience, invocation timing, and review scope are clear for pre-merge and code-change reviews. Non-fit boundaries, input/output expectations, semantic trigger exclusions, and Chinese or cross-agent environment support are not clearly specified, so points are deducted.
The document has clear sections for overview, triggers, process, categorization, checklist, verification, and related references. Repository context supplies an MIT license, maintainer information, and contribution path, but versioning, changelog, detailed maintenance ownership, installation notes, and FAQs are incomplete.
The skill provides directly usable five-axis review guidance, sizing and splitting strategies, severity labels, and verification requirements, so it can perform the core task. No representative review outputs, executed results, or comparative-benefit evidence are provided, leaving substantial judgment and tooling work.
The skill contains auditable checklists and verification criteria, and repository context includes CI validation. The supplied test fixture is unrelated to this skill, and there is no evidence independently reproducing its review behavior or quality, so the score is limited.
- This is a static source-only assessment; the skill, CI, and review process were not executed.
- The skill requires test and build verification but does not define concrete commands, output formats, or failure-handling protocols.
- Referenced security-and-hardening and performance-optimization materials were not supplied as selected evidence, so their completeness cannot be confirmed.
- The publisher is not verified by the FollowSkills enterprise registry and is treated as unknown.
What does this skill do, and when should you use it?
This skill defines a multi-axis code review workflow for AI coding agents and engineering teams. It evaluates correctness, readability and simplicity, architecture, security, and performance before merge, while requiring evidence from tests, builds, and other verification. It also covers severity labels, structural remedies, change sizing, dead-code hygiene, dependency upgrades, and review speed. It is a strong fit for teams seeking a consistent quality gate, but it does not replace project-specific testing or security tooling.
Establishes the change intent, specification, and expected behavior; reviews tests first for behavioral coverage, edge cases, and regression protection; walks through each changed file across five review axes; labels findings as Required, Critical, Nit, Optional, or FYI; checks the author's test, build, manual, and UI-verification evidence; identifies potentially orphaned code after refactoring and asks before deleting it; reviews dependency additions and upgrades through maintenance, vulnerability, license, changelog, lockfile, and test-verification checks.
- A developer has completed a feature or bug fix and needs a structured review before merging it.
- A team is evaluating code written by another agent or model and needs evidence-based correctness, security, and architecture feedback.
- A maintainer is reviewing a refactor and needs to determine whether it actually reduces complexity.
- A pull request contains a dependency upgrade or lockfile change and needs focused supply-chain and verification review.
- A change is too large to review comfortably and the team needs concrete strategies for splitting it into smaller units.
What are this skill's strengths and limitations?
- Defines five explicit review dimensions spanning behavior, structure, security, and performance.
- Requires reviewing tests and the verification story instead of treating passing tests as sufficient.
- Provides severity labels and named structural remedies that make feedback actionable.
- Includes practical guidance for change sizing, splitting, dead code, review cadence, and dependency upgrades.
- The SKILL.md does not specify commands for automatically running tests, builds, npm audit, or performance profiling.
- Review quality depends on the agent's access to the changed files, tests, and verification evidence.
- Detailed security and performance guidance points to other repository skills or reference files and is not fully contained in this one file.
- The source provides no standalone test suite, benchmark results, or platform compatibility test results.
How do you install this skill?
Install the individual skill with the repository's documented skills CLI command:npx skills add addyosmani/agent-skills --skill code-review-and-quality
The README does not document a single platform-neutral destination directory; alternatively, place the skill folder using the installation method supported by the target Agent Skills-compatible client.
How do you use this skill?
Trigger it before merging a change, for example: Review this change for correctness, readability, architecture, security, and performance before merge. It is also intended for completed features, refactors, bug fixes, and changes produced by another agent or model.
How does this skill compare with similar options?
The README names Superpowers and Matt Pocock's skills as alternatives and links to a repository comparison document, but the supplied source does not provide specific comparison conclusions.