From enpitech
Run a 5-pass language-agnostic code review scoped to the current PR diff. Use when reviewing pull requests for any language or framework that doesn't have a dedicated skill (Vue.js, Go, Rust, Ruby, PHP, Java, etc.).
How this skill is triggered — by the user, by Claude, or both
Slash command
/enpitech:cr-generalThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Review scope: **PR diff + directly affected files only.**
Review scope: PR diff + directly affected files only.
Follow the review criteria defined in rules/general.md.
Auto-detect the language and framework by scanning:
.vue, .go, .rs, .rb, .php, .java, .kt, .swift, .cs, etc.)package.json, go.mod, Cargo.toml, Gemfile, composer.json, build.gradle, pom.xml, pubspec.yaml, etc.State the detected language and framework at the top of your output.
Run these commands sequentially (no nested substitution):
gh pr view --json number,title,body,baseRefName,headRefName,url — store the baseRefName valuegit diff <baseRefName>...HEAD — using the baseRefName from step 1git diff --name-only <baseRefName>...HEAD — changed files listFor each changed file, identify direct consumers (files that import from it) to catch breakage from interface changes. Only expand one level deep — do not crawl the entire dependency graph.
Apply all 5 review passes and filtering rules from the criteria file against the diff and affected files. Adapt the checks to the detected language and framework.
Collect all findings with this format per issue:
### [SEVERITY] Pass N — file:line
**Issue:** one-line description
**Why:** reason you're confident this is a real issue
**Fix:** suggested code change
If no issues found: "✅ No issues found — all 5 review passes came back clean."
Show the user the findings.
Follow the autofix workflow defined in rules/autofix.md.
Local mode: present the three options (create findings file as cr-general-findings.md, fix step by step, fix all)
CI mode: post findings as PR comments with /fix and /fix-all reply instructions
Do not flag issues that are clearly intentional based on git blame context.
Run everything autonomously without asking to confirm each step (except the autofix choice in local mode).
Guides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.
npx claudepluginhub enpitech/ai-tools --plugin enpitech