Agent Routing: docs/
March 5, 2026 · View on GitHub
This file tells AI agents which subdirectory to search based on their intent.
Cursor Cloud agents: For build, test, and Cursor-specific constraints, read the root AGENTS.md first.
Intent → Directory map
| Intent | Directory | Notes |
|---|---|---|
| "How does the system work?" | design/architecture/ | Start with cursor-native-system-design.md |
| "How does each module work?" | design/architecture/cursor-drive-walkthrough.md | |
| "Why was this decision made?" | architecture/adr/ | ADRs 0001-0006; see adr/README.md |
| "What is the architecture?" | architecture/ | README + ADRs |
| "What are the requirements?" | prd/ | 5 PRDs covering all feature areas |
| "What are the requirements traceability?" | plans/ | traceability-matrix.md maps PRDs → plan TODOs |
| "What features are implemented?" | design/ux/ | features-implemented-flow.md, features-implemented-data-flow.md |
| "ADR validation?" | plans/ | adr-validation-map.md maps ADRs → tests/checks |
| "How do I set up / run?" | guides/getting-started.md | |
| "How do I test?" | guides/live-testing.md | |
| "What config key controls X?" | reference/config-schema.md | |
| "What MCP tools exist?" | reference/mcp-tools.md | |
| "What commands/keybindings exist?" | reference/commands-and-shortcuts.md | |
| "How does the voice pipeline work?" | prd/prd-voice-io.md + design/ai/prompt-optimizer-design.md | |
| "How does model selection work?" | design/ai/model-cost-tiers.md | |
| "How do agents work?" | prd/prd-multi-agent.md + design/architecture/cursor-native-system-design.md | |
| "What's the safety model?" | prd/prd-safety-config.md | |
| "How does the persona work?" | prd/prd-session-persona.md | |
| "What are the UX principles?" | design/ux/drive-mode-analysis-and-ux.md | |
| "What's the pair-programming philosophy?" | design/philosophy/senior-engineer-pair-programming.md | Steering, teaching, concise-first; ADR-0015 |
| "What's the user journey?" | design/ux/drive-mode-user-journey.md | |
| "Naming conventions?" | design/naming/cursor-aligned-naming.md | |
| "Continuing in a new chat?" | guides/handoff.md | |
| "Research / background reading?" | research/ | |
| "Voice integration options?" | research/openclaw-integration-analysis.md | |
| "Plan generation prompts?" | prompts/ |
Discovery protocol
When you need to understand an unfamiliar area:
- Read the relevant directory's
README.mdfirst — it summarizes all files. - Then read the specific file(s) it points to.
- For cross-cutting concerns (e.g., "how does config affect TTS?"), read
reference/config-schema.mdfor the config side andprd/prd-voice-io.mdfor the feature side.
What lives outside docs/
| What | Where |
|---|---|
| Executable plans | .cursor/plans/*.plan.md |
| Architecture overview | docs/architecture/README.md |
| Extension source | src/ |
| Plugin skills | .cursor/skills/ |
| Plugin rules | .cursor/rules/ |
| Plugin commands | .cursor/commands/ |
| Hooks | .cursor/hooks/ |