A comprehensive collection of 84 specialized AI subagents for Claude Code, organized by domain expertise.
All agents are organized into folders for easy navigation:
📐 Architecture & System Design - 12 agents
Design systems, build applications, and architect solutions.
- Core: backend-architect, frontend-developer, graphql-architect, architect-review
- Cloud: cloud-architect, hybrid-cloud-architect, kubernetes-architect
- UI/Mobile: ui-ux-designer, mobile-developer, ios-developer, flutter-expert, ui-visual-validator
💻 Programming Languages - 16 agents
Language-specific expertise from systems to web development.
- Systems: c-pro, cpp-pro, rust-pro, golang-pro
- Web: javascript-pro, typescript-pro, python-pro, ruby-pro, php-pro, sql-pro
- Enterprise: java-pro, scala-pro, csharp-pro
- Frameworks: elixir-pro, django-pro, fastapi-pro
🏗️ Infrastructure & Operations - 10 agents
DevOps, databases, and site reliability engineering.
- DevOps: devops-troubleshooter, deployment-engineer, terraform-specialist, dx-optimizer
- Databases: database-optimizer, database-admin
- SRE: incident-responder ⚡, observability-engineer, performance-engineer, network-engineer
🛡️ Quality & Security - 9 agents
Code review, security, and comprehensive testing.
- Review: code-reviewer, debugger, error-detective
- Security: security-auditor, backend-security-coder, frontend-security-coder, mobile-security-coder
- Testing: test-automator, tdd-orchestrator
🤖 Data & AI - 6 agents
Complete ML/AI lifecycle from research to production. All Opus models.
- data-scientist, data-engineer
- ai-engineer, ml-engineer, mlops-engineer
- prompt-engineer
📝 Documentation & Content - 6 agents
Technical writing and content creation.
- docs-architect, api-documenter, tutorial-engineer
- reference-builder, mermaid-expert, content-marketer
💼 Business & Operations - 7 agents
Analytics, finance, HR, legal, and customer ops.
- Analysis: business-analyst, quant-analyst, risk-manager
- Sales/Support: sales-automator, customer-support
- Compliance: hr-pro, legal-advisor
🎯 SEO & Content Optimization - 10 agents
Highest capability density in the entire collection.
- Complete SEO workflow from audit to authority building
- All agents: content-auditor, meta-optimizer, keyword-strategist, structure-architect, snippet-hunter, content-refresher, cannibalization-detector, authority-builder, content-writer, content-planner
⚙️ Specialized Domains - 8 agents
Unique platforms and use cases.
- blockchain-developer, payment-integration, legacy-modernizer
- context-manager, unity-developer, minecraft-bukkit-pro, quasar-developer, search-specialist
cd ~/.claude
git clone https://github.com/wshobson/agents.gitAgents are automatically available in Claude Code once placed in ~/.claude/agents/.
Automatic: Claude Code selects agents based on context
"Build a React dashboard with authentication"
→ Auto-activates: frontend-developer, backend-architect
Explicit: Specify agents by name
"Use security-auditor to scan for vulnerabilities"
"Have prompt-engineer optimize my LLM prompts"
| Model | Count | Use Case |
|---|---|---|
| Haiku | 12 | Quick, focused tasks (SEO, context, search) |
| Sonnet | 47 | Standard development and engineering |
| Opus | 26 | Complex reasoning, architecture, critical analysis |
Special Note:
- Only 1 agent uses IMMEDIATE activation:
incident-responder⚡ (for production crises) - 80 agents are PROACTIVE (designed to activate automatically)
- CATALOG.md - User-friendly quick reference by use case, skill level, and project phase
- AGENT-MATRIX.md - Visual ecosystem map with workflows and relationships
- catalog-analysis.py - Advanced analysis tool for agent metrics
Each category folder contains:
- Complete agent list with descriptions
- Common workflow patterns
- Use case selection guide
- Integration points
- Power combinations
architect-review → backend-architect → frontend-developer → test-automator → security-auditor → deployment-engineer
incident-responder ⚡ → devops-troubleshooter → observability-engineer
prompt-engineer → ai-engineer → ml-engineer → mlops-engineer
security-auditor → [backend + frontend + mobile]-security-coder → code-reviewer
cloud-architect → terraform-specialist → kubernetes-architect → observability-engineer
- prompt-engineer (250 lines, 150 capabilities) - LLM optimization master
- observability-engineer (210 lines, 156 capabilities) - Production monitoring expert
- unity-developer (207 lines, 148 capabilities) - Game development specialist
- test-automator (203 lines, 143 capabilities) - Quality assurance champion
- mlops-engineer (197 lines, 140 capabilities) - ML infrastructure architect
- seo-authority-builder - 59 capabilities/section
- seo-content-refresher - 48 capabilities/section
- seo-cannibalization-detector - 43 capabilities/section
- incident-responder ⚡ - Only IMMEDIATE activation agent
- security-auditor - Security gateway
- code-reviewer - Quality gateway
- observability-engineer - Production visibility
- database-optimizer - Performance critical
| Combo | Agents | Purpose |
|---|---|---|
| Security Gate | code-reviewer + security-auditor | Pre-deployment check |
| AI Dream Team | prompt-engineer + ai-engineer + mlops-engineer | LLM to production |
| Performance | performance-engineer + database-optimizer | Eliminate bottlenecks |
| Infrastructure | cloud-architect + terraform-specialist | Cloud automation |
| Quality First | tdd-orchestrator + test-automator | TDD workflow |
| Documentation | docs-architect + api-documenter + tutorial-engineer | Complete docs |
| SEO Campaign | Full SEO folder (10 agents) | Complete SEO strategy |
Beginner (Low complexity, quick wins):
- sales-automator, debugger, error-detective, mermaid-expert, reference-builder
Intermediate (Standard development):
- frontend-developer, api-documenter, tutorial-engineer, business-analyst, customer-support
Advanced (Deep expertise):
- backend-architect, cloud-architect, data-scientist, kubernetes-architect, test-automator
Expert (Maximum complexity):
- observability-engineer, prompt-engineer, mlops-engineer, incident-responder, security-auditor
- Kubernetes: 24 agents
- Databases: 21 agents
- AWS: 20 agents
- React: 16 agents
- Monitoring: 14 agents
- GCP: 13 agents
- Testing: 11 agents
- Azure: 10 agents
- Python ML: 7 agents
- Claude Code Documentation
- Subagents Documentation
- Claude Code Commands - 52 pre-built slash commands
- TDD Usage Examples
See CONTRIBUTING.md for guidelines on:
- Adding new agents
- Improving existing agents
- Quality standards
- Community guidelines
MIT License - see LICENSE file for details.
Need to...
- Build an API? →
agents/architecture/backend-architect.md - Fix production issue? →
agents/infrastructure/incident-responder.md⚡ - Optimize ML model? →
agents/data-ai/ml-engineer.md - Security review? →
agents/quality-security/security-auditor.md - Write documentation? →
agents/documentation/docs-architect.md - Improve SEO? →
agents/seo/(entire folder) - Modernize legacy code? →
agents/specialized/legacy-modernizer.md
84 Agents • 9 Categories • Production Ready