Route code diffs to specialized subagents for correctness, security, and performance review.
npx skills add https://github.com/kazukitcy/skills --skill review-code-kazukitcy
Please help me install this Agent Skill. Skill: review-code Source: https://github.com/kazukitcy/skills/tree/main/review-code Command: npx skills add https://github.com/kazukitcy/skills --skill review-code-kazukitcy
This skill solves the problem of generic, low-signal code reviews by orchestrating a specialized, risk-aware analysis that routes code changes to specific review lenses based on the nature of the diff.
Use the review-code skill to perform a comprehensive risk-routed review of the current staged changes in the repository.
High-intent search queries and answers about installing and using this skill.
To perform a multi-lens automated code review, you can orchestrate risk-routing of pull request diffs to specialized subagents. This process analyzes correctness, security, and performance independently against a shared rubric to identify high-signal findings.
Risk-routed code review is an automation approach that selects specific review lenses based on the risks identified in code changes. It improves quality assurance by deploying dedicated subagents for security, performance, and reliability, yielding severity-ranked actionable feedback.
Yes, you can apply automated static analysis to working tree changes, commits, and pull requests. The review skill evaluates diffs using specialized lenses for correctness and adversarial threats, ensuring architectural risks are identified before integration.
Yes, running orchestrated code review subagents requires a structured environment capable of spawning independent subagents. This setup is necessary to execute the individual review lenses against the shared rubric and aggregate consolidated findings.
The best way to consolidate security and performance findings is by aggregating results from specialized analysis lenses into a single severity-ranked report. This orchestration ensures high-signal feedback rather than generic approval by evaluating each risk area independently.
Automated code review returns low-signal findings when it applies generic analysis instead of risk-routing diffs to specialized subagents. By orchestrating dedicated lenses for security, performance, and correctness against a shared rubric, the review yields high-signal architectural risk identification.
GLaDO8Generates concise Git commit messages and commits staged or unstaged changes.
fyrsmithlabsCoordinate parallel research agents and synthesize findings into a consolidated report.
LuisSambranoPerform systematic code review and static analysis across multiple programming languages.
ybbms777Construct concrete failure scenarios for high-risk code diffs and output structured JSON findings.
