Defines the standard output format for all analysis agent findings. Consistent structure enables validation agents and consolidation to process findings uniformly.
Defines the fundamental principles that guide all Argus analysis. Every agent must internalize these principles to ensure consistent, high-quality recommendations.
Defines the standard output format for validation agent results. Links back to original findings and provides evidence for decisions.
Defines the contract for all validation agents. Validation agents verify findings from analysis agents, ensuring only proven issues reach the final report.
<!-- @include: ../_shared/philosophy.md -->
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
The all-seeing code reviewer
Argus is a multi-agent code review system designed to detect and prevent architecture drift in evolving codebases.
It runs specialized agents in parallel—covering areas like architecture, naming, dependency injection, dead code, and error handling—and validates every finding against the code. Argus also integrates with the Codex CLI and Gemini to cross-validate results across multiple LLMs, discarding any unverified issues and surfacing only accurate, actionable feedback.
# Add marketplace
/plugin marketplace add josstei/argus-claude
# Install
/plugin install argus@josstei-argus-claude
# Run a review
/argus:review # Shows confirmation prompt
/argus:review --yes # Accepts defaults immediately
Select a review level and go. No configuration needed.
| Level | Model | Time | Token Usage | Best For |
|---|---|---|---|---|
| Fast | Haiku | ~2 min | Low | Quick checks during development |
| Balanced | Sonnet | ~5 min | Moderate | Regular code reviews |
| Comprehensive | Opus | ~8 min | Higher | PR reviews, releases |
/argus:review # Shows confirmation prompt
/argus:review --level fast # Skip prompt, use Fast
/argus:review --level balanced # Skip prompt, use Balanced
/argus:review --yes # Skip prompt, use defaults
| Command | Description |
|---|---|
/argus:review | Run code review |
/argus:config | View settings |
/argus:help | Show help |
/argus:config set scope src/ # Always review src/
/argus:config set output markdown # Always save reports
Argus reads only the user-level config file:
~/.claude/argus.json (personal, all projects)| Phase | Focus |
|---|---|
| Architecture | Boundary violations, layer issues |
| Code Hygiene | Dead code, unused exports |
| Naming | Convention compliance |
| Quality | DI patterns, error handling, resources |
| Optimization | Performance opportunities |
Phases run based on depth setting:
quick - Architecture + Hygienestandard - All except Optimizationthorough - All phasesEnable cross-validation with other LLM CLIs:
/argus:review --external
Requires Codex CLI and/or Gemini CLI installed. Argus auto-detects availability and only enables detected tools.
Argus prioritizes long-term excellence over quick fixes:
MIT
npx claudepluginhub josstei/argus-claude --plugin argusMulti-agent development orchestration platform — 22 specialists, 4-phase orchestration, native parallel subagents, persistent sessions, and standalone review/debug/security/perf/seo/a11y/compliance commands
Reflective code review toolkit with structured analysis, blast radius analysis, domain expertise, and intelligent review orchestration
Multi-perspective code analysis covering architecture, security, and best practices
[Beta] 6つの専門エージェントチームでコードベースのアーキテクチャを多角的に分析するスキル
Comprehensive code review skill with 53 specialized agents for architecture, code quality, error handling, types, comments, tests, accessibility, localization, concurrency, performance, simplification, security, and platform-specific reviews (iOS, macOS, Android, Angular, TypeScript, Next.js, Vue.js, Python, Django, Ruby, Rust, Go, Rails, Flutter, Java/Spring Boot, C#/.NET, PHP/Laravel, C/C++, React Native, Svelte/SvelteKit, Elixir/Phoenix, Kotlin Server, Scala, Terraform, Shell/Bash, Docker, Kubernetes, GraphQL, GitHub Actions) with automatic platform detection and confidence scoring.
Live codebase visualization and structural quality gate — 14 health dimensions graded A-F, dependency analysis, and architecture governance via MCP
Automated code review with severity levels and actionable feedback