← ClaudeAtlas

analyzelisted

Analyze a project's codebase, PR review patterns, and conventions. Run this first before creating agents. Outputs a structured analysis that create-team uses.
Shaylyntriggerhappy850/claude-army · ★ 1 · AI & Automation · score 71
Install: claude install-skill Shaylyntriggerhappy850/claude-army
Analyze the current project to prepare for agent team creation. Read everything relevant and produce a structured analysis. ## Step 1: Project Foundation Read these files if they exist: - README.md, CONTRIBUTING.md, CHANGELOG.md - .github/pull_request_template.md - .github/ISSUE_TEMPLATE/ ## Step 2: Tech Stack Identify: - Language and version (go.mod, package.json, Cargo.toml, pyproject.toml, etc.) - Build system and commands - Linter config with exact rules - Test framework, assertion library, coverage requirements - CI/CD workflows (.github/workflows/) ## Step 3: Code Patterns - Map directory structure with Glob - Read 5-10 key source files for patterns - Read 5-10 test files for testing conventions - Identify the extension/plugin/registration pattern (how new features are added) ## Step 4: PR Review Analysis Find the lead reviewer: ```bash gh pr list --state all --limit 100 --json number,reviews --jq '.[].reviews[].author.login' 2>/dev/null | sort | uniq -c | sort -rn | head -5 ``` Find external contributor PRs: ```bash gh pr list --state all --limit 200 --json number,title,state,author --jq '.[] | "\(.number) [\(.state)] \(.author.login): \(.title)"' 2>/dev/null ``` For each external PR with lead reviewer feedback, extract BOTH: PR-level reviews: ```bash gh api repos/OWNER/REPO/pulls/NUMBER/reviews --jq '.[] | select(.user.login == "REVIEWER") | .body' 2>/dev/null ``` Inline code comments: ```bash gh api repos/OWNER/REPO/pulls/NUMBER/comments --jq '.[] | selec