Model Context Protocol) server for AI-assisted scheduler management.
## Essential Build Commands
### Building the MCPServer and CLI
```bash
# Build the schedcp MCPserver and CLI tool
cd mcp
cargo
/generate-image` for reference images
2. Use `/generate-video` for video clips
3. Use ElevenLabs MCP tools for voiceovers
4. Combine assets in Remotion compositions
### Timed Content from Transcripts
1. `/transcribe video.mp4
teams are defined in `templates/.claude/agents/`. See `templates/CLAUDE.md` for the canonical campaign workflow.
| Agent | Role | MCPServers |
|-------|------|-------------|
| Research | Target analysis, literature, prior art | pdb, uniprot, sabdab, research |
| Design | Generate designs
load a specific tool,
or `"+by-pdb"` to find all tools from a server.
Use MCP research tools FIRST. Never default to web search when structured databases are available.
**Servers
have three interfaces:
1. **MCP tools** (`video_editor_*`) — live timeline control via the MCPserver
2. **CLI** (`editorctl`) — direct commands from this terminal
3. **HTTP API** — POST to `http://localhost
context.
Show the results to the user.
### 2. DURING THE SESSION — automatic saving
The MCPserver automatically saves each interaction in sessions/ .jsonl.
You don't need to do anything
commits
git push
```
## Debugging backend data (MCP)
When the backend is running, the `moqui` MCPserver exposes `moqui_rest_call` — a read-only
(GET) tool to inspect live data
suite** - ALL tests in the project must pass
- [ ] **Full E2E test with chrome-devtools MCP** - Test complete user flows
- [ ] **Build succeeds** - `npm run build` must pass
- [ ] Verify the entire application
CLAUDE.md files**: When adding, removing, or changing the responsibility of commands, services, templates, agents, MCPservers, types, or migrations, update the CLAUDE.md in the corresponding directory. See [Component Documentation](#component
permanent notes
- Focus on unique/contrarian perspectives
- Return text suitable for content use
---
**Integration with MCPServers:**
Your environment includes MCPservers that provide additional capabilities:
**1. Local Brain Search - Memory
named `mcp__al-symbols-mcp__*`, `mcp__context7__*`, `mcp__microsoft-docs__*` (or the `mcp__plugin_aldc_…` form). A server that is not loaded is `unavailable`; do not simulate it |
| `sshadowsdk.al
status
# Run diagnostics
./ckb doctor
# Start HTTP API server
./ckb serve --port 8080
# Start MCPserver (for AI tool integration)
./ckb mcp
# Start A2A protocol server (for agent-to-agent
engineer collaborating on `crawler_agent`. Your primary goal is to expose Crawl4AI via an MCPserver that integrates cleanly with OpenAI Agents SDK and works well in Cursor/Cloud Code.
## Mission
validation (lint/build) → commit → repeat or exit.
**LLM Providers** - For internal features (Anthropic, OpenAI, OpenRouter)
**MCPServer** - Model Context Protocol server exposing ralph-starter capabilities as tools (init, plan, run)
## Important
fall back to ezdxf
3. On non-Windows: always use ezdxf
### FastMCP Server Structure (`server.py`)
The `mcp` FastMCP instance is configured with:
- **Lifespan** (`autocad_lifespan`): initializes the backend singleton; stores
A file Claude Code reads at the start of every session. It holds the commands, conventions and warnings the agent needs for this project.
Where does it go?
At the repository root. Claude Code also reads CLAUDE.md files in subdirectories when it works there.
What should it contain?
Build and test commands, the project's layout, conventions that aren't obvious from the code, and mistakes to avoid. Short files tend to work better than long ones.
CLAUDE.md or AGENTS.md?
Claude Code reads CLAUDE.md; most other agents read AGENTS.md. Many projects keep one and point the other at it.