This directory contains technical implementation details and design docs for Resonnet.
| Doc | Description |
|---|---|
| architecture.md | Overall architecture: API layer, agent layer, workspace layout, data flow |
| assignable-skills-flow.md | Assignable skills API, copy logic, meta aggregation |
| config.md | Environment variables; .env loading order; ANTHROPIC_* vs AI_GENERATION_*; Profile Helper auth modes (none/jwt/proxy) |
| mcp-config.md | MCP config API, validation (npm/uvx/remote only, no local) |
| testing.md | Test layers (unit/integration), .env setup, CI notes |
| api-reference.md | API endpoint list and brief descriptions |
| runtime-modes.md | How to use Resonnet in executor mode vs standalone MVP mode |
| agent-links-runtime.md | Agent link runtime lifecycle: per-session workspace, SSE chat, import limits |
| sandbox-isolation.md | Topic-level sandbox design, OS isolation model, and srt integration details |
| skills-generalization.md | Library design; experts in libs/experts/; discussion modes in libs/moderator_modes/ |
| skills-submodule-guide.md | Add/update skill libraries via submodule; points to Cursor skill |
| import-skill-repo.md | One-click import script for external skill repos |
| troubleshooting.md | Operational troubleshooting for install, runtime, and environment issues |
| ../app/prompts/README.md | AI prompt management (expert/moderator generation, round discussion, expert reply) |
| ../libs/README.md | Libs directory structure; experts, moderator_modes, mcps, prompts |
- Getting started: Read config.md for env setup, then testing.md to run tests
- Architecture: architecture.md
- Runtime mode selection: runtime-modes.md
- API development: api-reference.md
- AgentSDK availability must be validated with a real
.env;ANTHROPIC_API_KEY=testis not acceptable for acceptance. - In TopicLab integrated mode, topic/domain business storage belongs to
topiclab-backend; Resonnet acts as the agent execution backend.