{ "boardId": "your-board-id", "apiKey": "your-api-key", "apiToken": "your-api-token", "workflow": { "todo": ["list-id-for-todo"], "inProgress": ["list-id-for-in-progress"], "review": ["list-id-for-review"], "done": ["list-id-for-done"] } } ``` Notes: - `workflow` keys are arbitrary — `todo`, `inProgress`, `review`, and `done` are the ones that default to `[]` if you omit them, but you can add any other stage name (e.g. `testing`, `validated`, `rejected`) and it will be picked up as-is. - Each stage maps to an array of list IDs, so multiple lists can share one stage if needed. Any file matching `*.config.board.json` is automatically gitignored — this is where your credentials live, so never commit it or rename it out of that pattern. Point the CLI at your config in one of three ways, in priority order: ```bash # 1. --config-file flag (highest priority) trello tasks list --config-file ./myboard.config.board.json # 2. BOARD_CONFIG environment variable export BOARD_CONFIG=./myboard.config.board.json trello tasks list # 3. .env file (loaded automatically via dotenv) echo "BOARD_CONFIG=./myboard.config.board.json" > .env trello tasks list # 4. Falls back to ./trello.config.board.json in the current directory if none of the above is set ``` A `.env.example` is included as a template for the `.env` approach. ## CLI Usage ### Lists ```bash trello lists # List all lists with IDs and card counts ``` ### Tasks (multi-card) ```bash trello tasks list # List all tasks (JSON, auto-cached) trello tasks list --stage todo # Filter by workflow stage trello tasks list --list "BackEnd" # Filter by list name trello tasks list --list-id # Filter by list ID trello tasks list --limit 10 # Limit results trello tasks list --brief # Concise output (id, name, list, labels) trello tasks list --format table # table | text | json (default) trello tasks move --to "testing" # Move specific cards to a list trello tasks move --from "BackEnd" --to "done" # Move all cards from one list to another ``` ### Task (single-card) ```bash trello task get # Full context: card, comments, attachments, checklists trello task get --no-comments trello task get --no-attachments trello task get --no-checklists trello task get --brief # Concise output (id, name, desc, list, labels, comment text) trello task move

# Move by list ID, workflow stage, or list name trello task comment "message text" trello task comment --file comment.txt # Read comment from a file trello task comment --stdin # Read comment from stdin (pipe) — both avoid shell-escaping issues trello task label

# Add a label (matches by ID, name, or color) trello task label

--remove ``` ### Labels ```bash trello labels list # List all board labels ``` ### Attachments ```bash trello attachments list trello attachments download trello attachments download --output ./my-attachments ``` ### Global Options ```bash --config-file # Override config file --format # Output format: json (default), table, text --debug # Enable debug logging ``` **Tips for scripting / AI agents:** - Use `--brief` to reduce token usage — it strips timestamps, URLs, and member IDs. - Use `--format json` (the default) for structured parsing. - Run `trello lists` first to discover list names/IDs before filtering by them. - Use `--file`/`--stdin` for multiline comments instead of trying to escape them on the command line. ## MCP Server (wiring into Claude Code, Cursor, etc.) The project ships an MCP server (`src/mcp.ts`) that exposes every Trello operation as a tool over JSON-RPC 2.0 stdio — no extra MCP SDK dependency required. ```bash yarn dev:mcp # run in dev mode node dist/mcp.js # run the built server ``` Wire it into an MCP-compatible client by pointing it at the built server and passing your board config via `BOARD_CONFIG`: ```json { "mcpServers": { "trello": { "command": "node", "args": ["/dist/mcp.js"], "env": { "BOARD_CONFIG": "/myboard.config.board.json" } } } }

Trello MCP Server

Installation