Audit code diffs for regression risks and breaking changes.
npx skills add https://github.com/Virgile-Eratel/skills --skill lverif
Please help me install this Agent Skill. Skill: lverif Source: https://github.com/Virgile-Eratel/skills/tree/main/lverif Command: npx skills add https://github.com/Virgile-Eratel/skills --skill lverif
This Skill reduces the chance you ship regressions after development by quickly auditing changes and enumerating the most likely risks.
Ask the AI to run the lverif audit automatically after the plan is finished to get a prioritized list of concrete risks from your diff.
High-intent search queries and answers about installing and using this skill.
A post-development code diff audit reviews your working changes and branch-vs-main differences to identify regression risks, breaking API changes, typing breakages, and security gaps. It generates a prioritized P0/P1/P2 report with specific file:line anchors for each detected risk.
Regression risk detection works by comparing your current branch against main using git diff main...HEAD. It evaluates modified exported interfaces, routes, and schemas to enumerate potential type issues and unhandled edge cases, outputting a low-noise prioritized risk report.
Yes, you can automate a security and risk audit to run after development finishes. The workflow can auto-run the audit immediately after plan completion, analyzing your git diff to enforce read-only behavior and flag security gaps without requiring manual intervention.
Yes, the risk audit supports deep analysis for exported schemas and routes. It automatically switches from a fast audit mode to a deeper audit mode when exported interfaces, routes, or schemas are modified, ensuring comprehensive breaking change detection.
The best way to generate a prioritized P0/P1/P2 report for code review is to run an automated post-dev audit on your git diff. This enforces read-only behavior and outputs a low-noise report with file:line evidence for each identified regression risk and security gap.
Your code review audit outputs in French because the tool enforces strict French-first output by default. This design choice ensures the prioritized risk reporting aligns with its intended low-noise, strict formatting standards for the target development environment.
chrisliu298Remove AI-generated commentary and dead code from git diffs.
huangwb8Analyze code changes and prioritize P0/P1/P2 issues before merging.
rahmatullahbossGenerate automated end-to-end test cases for existing platform features.
xotongSimulate DAST security testing via static codebase analysis without running the application.
