|
| 1 | +# Global Claude Instructions |
| 2 | + |
| 3 | +## Agent Dispatch Matrix |
| 4 | + |
| 5 | +Automatically invoke specialized agents based on task patterns and keywords. |
| 6 | + |
| 7 | +### Core Development Agents |
| 8 | + |
| 9 | +| Keywords | Agent | Use When | |
| 10 | +| -------------------------------------------------------------------------------- | --------------------------- | ----------------------------------- | |
| 11 | +| error, bug, crash, exception, stack trace, debug, not working | `debugger` | Something is broken, need diagnosis | |
| 12 | +| code review, PR, pull request, review code, check code, feedback | `code-reviewer` | Standard code review | |
| 13 | +| thorough review, find problems, aggressive review, validate claims, code quality | `adversarial-code-reviewer` | Need rigorous validation of work | |
| 14 | +| refactor, clean up, simplify, reduce complexity, technical debt, code smell | `refactoring-specialist` | Restructuring existing code | |
| 15 | +| write tests, test coverage, unit test, integration test, E2E test, automation | `test-automator` | Creating or improving tests | |
| 16 | + |
| 17 | +### Language-Specific Agents |
| 18 | + |
| 19 | +| Keywords | Agent | Use When | |
| 20 | +| ----------------------------------------------------------------- | ------------------- | ---------------------------- | |
| 21 | +| JavaScript, JS, ES2023, Node.js, vanilla JS, async | `javascript-pro` | JavaScript development | |
| 22 | +| TypeScript, TS, type safety, tsconfig, types | `typescript-pro` | TypeScript development | |
| 23 | +| Python, Python 3.11, data science, pandas, numpy | `python-pro` | Python development | |
| 24 | +| Rust, systems programming, memory safety, WebAssembly, Cargo | `rust-engineer` | Rust development | |
| 25 | +| React, React 18, hooks, component, Redux, Next.js, JSX | `react-specialist` | React development | |
| 26 | +| Svelte, SvelteKit, runes, $state, $derived, component, reactivity | `svelte-pro` | Svelte/SvelteKit development | |
| 27 | +| SQL, database, query, PostgreSQL, MySQL, optimization, index | `sql-pro` | Database/SQL work | |
| 28 | +| GraphQL, federation, subscriptions, Apollo, schema | `graphql-architect` | GraphQL API design | |
| 29 | + |
| 30 | +### Frontend & UI Agents |
| 31 | + |
| 32 | +| Keywords | Agent | Use When | |
| 33 | +| ------------------------------------------------------------------ | ---------------------- | --------------------- | |
| 34 | +| frontend, React, Vue, UI, component, CSS, responsive, styling | `frontend-developer` | Frontend development | |
| 35 | +| UI design, visual design, interface, interaction, design system | `ui-designer` | UI/visual design work | |
| 36 | +| UX research, user research, usability, user testing, user insights | `ux-researcher` | Research/UX analysis | |
| 37 | +| accessibility, WCAG, a11y, screen reader, inclusive design, ADA | `accessibility-tester` | Accessibility testing | |
| 38 | + |
| 39 | +### Backend & Infrastructure |
| 40 | + |
| 41 | +| Keywords | Agent | Use When | |
| 42 | +| ------------------------------------------------------------------- | --------------------- | ---------------------------- | |
| 43 | +| backend, server, API, endpoint, microservices, architecture | `backend-developer` | Server-side development | |
| 44 | +| fullstack, full-stack, end-to-end feature, database to UI | `fullstack-developer` | Complete feature development | |
| 45 | +| deploy, deployment, CI/CD, pipeline, release automation | `deployment-engineer` | Deployment & release | |
| 46 | +| DevOps, infrastructure, operations, container, Docker, Kubernetes | `devops-engineer` | Infrastructure & operations | |
| 47 | +| network, networking, cloud architecture, security, zero-trust | `network-engineer` | Network & infrastructure | |
| 48 | +| WebSocket, real-time, realtime, Socket.io, messaging, bidirectional | `websocket-engineer` | Real-time communication | |
| 49 | + |
| 50 | +### Security & Compliance |
| 51 | + |
| 52 | +| Keywords | Agent | Use When | |
| 53 | +| -------------------------------------------------------- | -------------------- | -------------------------- | |
| 54 | +| security, vulnerability, CVE, auth, injection, XSS, CSRF | `security-engineer` | Security concerns | |
| 55 | +| compliance, GDPR, HIPAA, PCI DSS, audit, regulation | `compliance-auditor` | Compliance/regulatory work | |
| 56 | + |
| 57 | +### Performance & Monitoring |
| 58 | + |
| 59 | +| Keywords | Agent | Use When | |
| 60 | +| ------------------------------------------------------------- | ---------------------- | ------------------------ | |
| 61 | +| slow, performance, optimize, latency, bottleneck, profiling | `performance-engineer` | Performance optimization | |
| 62 | +| monitoring, metrics, anomaly detection, observability, alerts | `performance-monitor` | System monitoring | |
| 63 | + |
| 64 | +### Documentation & Writing |
| 65 | + |
| 66 | +| Keywords | Agent | Use When | |
| 67 | +| --------------------------------------------------------- | ------------------------ | ------------------------ | |
| 68 | +| document, documentation, README, API docs, write docs | `documentation-engineer` | Technical documentation | |
| 69 | +| API documentation, OpenAPI, Swagger, write API docs | `api-documenter` | API documentation | |
| 70 | +| writing, technical writing, user guide, tutorial, content | `technical-writer` | Technical writing | |
| 71 | +| SEO, search, ranking, structured data, optimization | `seo-specialist` | SEO/content optimization | |
| 72 | + |
| 73 | +### API & Architecture |
| 74 | + |
| 75 | +| Keywords | Agent | Use When | |
| 76 | +| ---------------------------------------------------------- | -------------- | --------------- | |
| 77 | +| API design, REST, GraphQL, endpoints, schema, architecture | `api-designer` | API design work | |
| 78 | + |
| 79 | +### Build & Tools |
| 80 | + |
| 81 | +| Keywords | Agent | Use When | |
| 82 | +| ------------------------------------------------------------------- | ---------------------- | -------------------- | |
| 83 | +| build, compilation, build system, Webpack, Vite, build optimization | `build-engineer` | Build system issues | |
| 84 | +| CLI, command-line, terminal, tool development, arguments | `cli-developer` | CLI tool development | |
| 85 | +| git, branching, merge, version control, rebase, workflow | `git-workflow-manager` | Git/version control | |
| 86 | + |
| 87 | +### Prompt & LLM |
| 88 | + |
| 89 | +| Keywords | Agent | Use When | |
| 90 | +| --------------------------------------------------------------- | ----------------- | ----------------- | |
| 91 | +| prompt, LLM, optimize prompt, token, few-shot, chain-of-thought | `prompt-engineer` | LLM prompt design | |
| 92 | + |
| 93 | +### System & Resilience |
| 94 | + |
| 95 | +| Keywords | Agent | Use When | |
| 96 | +| ------------------------------------------------------------ | ----------------------- | ----------------------- | |
| 97 | +| chaos, resilience, failure injection, load test, stress test | `chaos-engineer` | Resilience testing | |
| 98 | +| workflow, orchestration, process automation, state machine | `workflow-orchestrator` | Workflow design | |
| 99 | +| error handling, failure recovery, error coordination | `error-coordinator` | Error handling strategy | |
| 100 | +| error analysis, root cause, error pattern, logs, correlation | `error-detective` | Error pattern analysis | |
| 101 | + |
| 102 | +### Multi-Agent Coordination |
| 103 | + |
| 104 | +| Keywords | Agent | Use When | |
| 105 | +| ------------------------------------------------------------------- | ------------------------- | ------------------------- | |
| 106 | +| multi-agent, agent orchestration, agent coordination, team assembly | `agent-organizer` | Multi-agent system design | |
| 107 | +| multi-agent coordination, parallel execution, dependencies | `multi-agent-coordinator` | Multi-agent orchestration | |
| 108 | +| knowledge synthesis, patterns, insights, collective learning | `knowledge-synthesizer` | Knowledge extraction | |
| 109 | +| context management, state, information sharing | `context-manager` | Context/state management | |
| 110 | + |
| 111 | +## Dispatch Rules |
| 112 | + |
| 113 | +1. **Keyword Matching**: Scan user request for trigger keywords |
| 114 | +2. **Invoke Agent**: When a clear match is found, invoke with: `Task` tool with `subagent_type` |
| 115 | +3. **Announce**: State which agent is being used and why |
| 116 | +4. **Override**: User explicit request always supersedes automatic dispatch |
| 117 | +5. **Ambiguity**: If multiple agents match, ask user which fits best OR use most specific match |
| 118 | +6. **Compound Tasks**: For multi-phase work, invoke sequential agents as needed |
| 119 | + |
| 120 | +## Examples |
| 121 | + |
| 122 | +User: "There's a bug in my React component" |
| 123 | +→ Triggers: "bug", "React" → Invoke: `debugger` + `react-specialist` |
| 124 | + |
| 125 | +User: "Make sure this code is production ready" |
| 126 | +→ Triggers: "production ready", validation → Invoke: `adversarial-code-reviewer` |
| 127 | + |
| 128 | +User: "Design a REST API for user management" |
| 129 | +→ Triggers: "design", "REST API" → Invoke: `api-designer` |
| 130 | + |
| 131 | +User: "Slow database queries need optimization" |
| 132 | +→ Triggers: "slow", "database" → Invoke: `performance-engineer` + `sql-pro` |
0 commit comments