Your agents run the board too
Runlane ships a built-in MCP server. Claude Code, Cursor, or any MCP client connects with a personal token and gets the full toolset — create, move, and close tasks under the same roles as human teammates.
Two steps to connect
1 · Get a personal token
Sign in and generate a token on the MCP page in your profile. The token acts with your permissions on your boards.
Sign in and get a token2 · Add the server to your client
For Claude Code it is one command (paste your token in place of <TOKEN>):
claude mcp add --transport http --scope user runlane-board https://runlane.run/api/mcp --header "Authorization: Bearer <TOKEN>"26 tools
Everything an agent needs to genuinely work on the board — not just read it:
The skill that teaches the workflow
The tools tell an agent what it can do. The skill tells it when. Drop this file into Claude Code and your agent starts treating the board as part of the job — not a thing it touches when reminded.
- Finds the board for the repo once, then remembers it
- Asks before creating a card: create, skip, or create and drive it to done
- Pulls from Backlog, parks in In progress, closes in Done — by column role, not by name
- Ticks your plan off as subtasks, so progress is visible while it works
Install it from the Runlane marketplace, then restart Claude Code:
claude plugin marketplace add https://runlane.run/marketplace.json && claude plugin install runlane-board@runlaneSame rules as people
An agent inherits the role of the person who minted the token: what you can do, it can do — nothing more. Every change lands on the board in real time, visible to the whole team.