code-analysis

Plugin code-analysis from marketplace jylee-claude-marketplace in JuyeongYi/jylee_claude_marketplace. Commands read from the repo's manifest — 1 month ago.

Skills (17)

• analyze-file

  Performs structure + deep analysis of a single source file. Adapts analysis focus based on file_role (declaration, implementation, module). For implementation files, reads paired declaration mirror as context. Produces a mirror document at .codeanalyzer/mirror/{path}.md.

• analyze-pr

  Run a PR-aware quality audit. Uses mcp__gitnexus__detect_changes + mcp__gitnexus__impact to identify affected subsystems, then runs Phase 2 quality-analyzer on that subset only. Fast (minutes). Use when the user wants to audit a pull request, working-tree changes, or staged changes before commit. Ac

• analyze

  Comprehensive source code analysis on top of GitNexus knowledge graph. Phase 1 (structure, declarations) → Phase 2 (implementation quality, per subsystem) → Phase 3 (deterministic flow + dead code via GitNexus) → Phase 4 (usage docs, optional). Supports glob and natural-language scope. Hard-requires

• api-reference

  Phase 4: Generate API reference document. Compiles the Exports/Public API sections from Phase 1 mirror documents and class hierarchy information into a public API listing with signatures and usage examples.

• architecture-overview

  Phase 4: Generate project architecture overview document. Synthesizes Phase 1 structural analysis (subsystems, dependencies, class hierarchy) and Phase 3 flow analysis (entry points, call graphs) into a comprehensive architecture description.

• detect-language

  Counts source files by extension within the given scope, identifies the dominant programming language, and asks the user to confirm. Returns the confirmed language and file count breakdown.

• entrypoint-guide

  Phase 4: Generate entry point guide document. Based on Phase 3 flow analysis including entry points, call graphs, and event loop boundaries, describes app startup flow and major execution paths.

• extension-guide

  Phase 4: Generate dependency/extension guide. Analyzes Phase 1 class hierarchy, override patterns, and dependency structure to guide which classes to inherit and where to connect when adding new features.

• next-phase

  Transition to the next analysis phase after completing the current one. Phase 1→2: implementation analysis after subsystem selection. Phase 2→3: execution flow and dead code analysis. Blocks transition via phase-state.json guards when prerequisites are not met.

• phase1-report

  Phase 1 structural analysis report generation. Produces architecture summary, hotspots, issue catalog, and recommendations from aggregate data. Generates subsystems.md (with machine-readable YAML frontmatter) to support automatic subsystem detection in Phase 2.

• phase2-report

  Phase 2 implementation quality report generation. Structured by subsystem sections. Highlights code duplication, God Functions, and actual bugs. Includes header-implementation mismatches by comparing against Phase 1 header mirrors.

• phase3-report

  Phase 3 flow/dead code report generation. Includes entry point listing, call graph summary, and dead code listing. Contains cross-phase synthesis analysis (synthesis.md) correlating Phase 1/2 results.

• phase3-trace

  Phase 3: 진입점부터 실행 흐름 추적, dead code 탐지, 호출 그래프 생성. mcp__codeanalyzer__get_phase_state 기반 가드로 Phase 1 미완료 시 진행 차단. /next-phase 또는 직접 실행 가능.

• phase4-report

  Generate Phase 4 comprehensive index. Reads the 5 usage documents (architecture-overview, api-reference, entrypoint-guide, subsystem-guide, extension-guide) and consolidates them into index.md.

• pr-report

  Assemble PR-specific report from pr-scope + affected-subsystem quality results. Invoked by /analyze-pr after quality-analyzer runs complete. Writes to .codeanalyzer/reports/pr/<ref>/.

• subsystem-classifier

  Cluster analyzed files into subsystems using GitNexus embeddings (cosine similarity). Output a subsystems.md with machine-readable YAML frontmatter and a human-readable table. Replaces the v1 dependency-cluster + naming-pattern heuristic for more robust classification on repos without strict naming

• subsystem-guide

  Phase 4: Generate per-subsystem usage guide. Based on Phase 2 analysis results (per-subsystem mirrors and reports), documents each subsystem's role, key classes, and usage methods.

MCP servers (1)

• codeanalyzer

  runs: codeanalyzer mcp

Hooks (2)

• SessionStart

  bash ${CLAUDE_PLUGIN_ROOT}/hooks/session-start.sh

• PostToolUse

  bash ${CLAUDE_PLUGIN_ROOT}/hooks/after-agent.sh

Similar plugins