$curl -o .claude/agents/README.md https://raw.githubusercontent.com/travisjneuman/.claude/HEAD/agents/README.mdSpecialized AI agents for focused, expert-level assistance across all domains.
| 1 | # Claude Code Subagents |
| 2 | |
| 3 | Specialized AI agents for focused, expert-level assistance across all domains. |
| 4 | |
| 5 | **67 agents available** - Covering code quality, development, architecture, languages, platforms, creative, product & strategy, SaaS, and infrastructure. |
| 6 | |
| 7 | --- |
| 8 | |
| 9 | ## Parallel Execution Capabilities |
| 10 | |
| 11 | | Capability | Limit | |
| 12 | | --------------------- | -------------------------------------------------- | |
| 13 | | **Concurrent agents** | Up to 10 simultaneous | |
| 14 | | **Context per agent** | Independent (up to 1M tokens with Opus 4.6) | |
| 15 | | **Nesting** | Single-level only (agents cannot spawn sub-agents) | |
| 16 | | **Overhead** | ~20k tokens per Task tool invocation | |
| 17 | |
| 18 | ### Parallel Execution Pattern |
| 19 | |
| 20 | ``` |
| 21 | Main Claude (Orchestrator) |
| 22 | ├── Task(code-reviewer) ─┐ |
| 23 | ├── Task(test-generator) │ Run simultaneously |
| 24 | ├── Task(security-auditor) ─┘ |
| 25 | └── Collect results when all complete |
| 26 | ``` |
| 27 | |
| 28 | ### When to Use Parallel Agents |
| 29 | |
| 30 | - **Code review + test generation** - Run code-reviewer and test-generator in parallel |
| 31 | - **Multi-platform builds** - Run ios-developer and android-developer simultaneously |
| 32 | - **Comprehensive audits** - Run security-auditor, performance-optimizer, and accessibility-expert together |
| 33 | - **Documentation blitz** - Run documentation-writer on multiple modules |
| 34 | |
| 35 | --- |
| 36 | |
| 37 | ## Agent Teams |
| 38 | |
| 39 | Agent Teams coordinate multiple full Claude Code sessions working in parallel, each with its own context window and inter-agent messaging. Unlike subagents (Task tool), teammates communicate directly with each other and persist independently. |
| 40 | |
| 41 | | Approach | Context | Communication | Token Cost | Best For | |
| 42 | | --------------- | ---------- | --------------------------- | ---------- | ---------------------------------- | |
| 43 | | **Subagents** | Own window | One-way (to caller) | Low-Medium | Focused tasks, parallel search | |
| 44 | | **Agent Teams** | Own window | Two-way (between teammates) | High | Research, review, cross-layer work | |
| 45 | |
| 46 | Use `/assemble-team` to quickly spawn a pre-built team, or describe the composition in natural language. |
| 47 | |
| 48 | See `~/.claude/docs/AGENT-TEAMS.md` for the full guide and `~/.claude/docs/reference/workflows/agent-teams.md` for the decision framework. |
| 49 | |
| 50 | --- |
| 51 | |
| 52 | ## Available Agents |
| 53 | |
| 54 | ### Code Quality |
| 55 | |
| 56 | | Agent | Purpose | Model | |
| 57 | | ------------------------ | ------------------------------------------ | ------ | |
| 58 | | `deep-code-reviewer` | Comprehensive 6-aspect code review | opus | |
| 59 | | `security-auditor` | OWASP Top 10 and security analysis | sonnet | |
| 60 | | `refactoring-specialist` | Safe, incremental code improvement | sonnet | |
| 61 | | `performance-optimizer` | Bottleneck identification and optimization | sonnet | |
| 62 | |
| 63 | ### Development |
| 64 | |
| 65 | | Agent | Purpose | Model | |
| 66 | | ---------------------------- | ----------------------------------------- | ------ | |
| 67 | | `test-generator` | TDD-focused test suite generation | sonnet | |
| 68 | | `e2e-test-specialist` | Playwright/Cypress E2E and visual testing | sonnet | |
| 69 | | `api-designer` | REST/GraphQL API design | sonnet | |
| 70 | | `api-integration-specialist` | Third-party APIs, OAuth flows, webhooks | sonnet | |
| 71 | | `database-expert` | Schema design, query optimization, ops | sonnet | |
| 72 | | `devops-engineer` | CI/CD, containers, cloud, IaC | sonnet | |
| 73 | |
| 74 | ### Architecture |
| 75 | |
| 76 | | Agent | Purpose | Model | |
| 77 | | -------------------------- | ------------------------------------------- | ------ | |
| 78 | | `architecture-analyst` | System design and pattern analysis | opus | |
| 79 | | `cloud-architect` | Multi-cloud, serverless, DR planning | opus | |
| 80 | | `debugging-specialist` | Systematic 4-phase debugging | sonnet | |
| 81 | | `devsecops-engineer` | CI/CD security, SAST/DAST, supply chain | sonnet | |
| 82 | | `graphql-architect` | GraphQL schema and resolver design | sonnet | |
| 83 | | `microservices-architect` | Distributed systems and service design | sonnet | |
| 84 | | `migration-specialist` | Framework upgrades, codebase migrations | opus | |
| 85 | | `observability-engineer` | OpenTelemetry, SLOs, distributed tracing | sonnet | |
| 86 | | `realtime-specialist` | WebSockets and live data systems | sonnet | |
| 87 | | `serverless-specialist` | Lambda, Workers, Edge Functions, WASI | opus | |
| 88 | | `event-driven-architect` | Kafka, RabbitMQ, event sourcing, CQRS | opus | |
| 89 | |
| 90 | ### Language & Framework Experts |
| 91 | |
| 92 | | Agent | Purpose | Model | |
| 93 | | ------------------------- | ---------------------------------- | ------ | |
| 94 | | `typescript-expert` | Type system mastery | sonnet | |
| 95 | | `pyth |