planka-trackinglisted

# planka-tracking — onboard a project onto a Planka board This skill brings the **current project** under board-based task tracking: one Planka project + board, a small set of lists and labels, and (if the project already has a .claude/progress.md buffer) a zero-loss migration of its tasks into cards. It runs *in the project's own session* — never on behalf of another project from a different session (the domain discovery and the zero-loss migration both need the project's own files and conversation loaded). > **Hard dependency: the `planka` MCP server.** This skill only works if a `planka` MCP server > is installed and reachable in the session — it is how every board/list/label/card is created, > there is no REST-credential fallback. If you don't have one set up, stop here and set one up > first (any MCP server that implements Planka's API against your own Planka instance). Don't > attempt a partial onboarding without it. ## Preconditions — check before doing anything 1. **Confirm the project is unmanaged**: no .claude/planka.json at the project root. If it exists, the project is already onboarded — stop and tell the user (offer to reconcile the board against the current .claude/progress.md/state instead of re-onboarding). 2. **Confirm the `planka` MCP is reachable** in this session. The tools are MCP tools — discover their exact names with `ToolSearch` (`query: "planka"`) before calling them; do not hardcode signatures, they vary by server implementation. If