Start a sprint with explicit git_dir and a worktree-scoped, uncommitted data_dir; returns orientation.

Resume an existing sprint by binding this MCP session to an existing git_dir and data_dir without creating a sprint.

Detach this MCP session from its current Sprinty binding and stop the dashboard.

List existing sprint ledgers in a data_dir, or the current binding when already attached.

Close the sprint after a programmatic re-check of all gates.

Archive the sprint with a recovery reason, bypassing normal close gates.

Compact sprint overview: title, details, notes, artifacts, and subsprint counts.

Compact work window for choosing the next item.

Create a subsprint (description, goals, gates).

Read one subsprint with compact item rows.

List subsprints with compact rows.

Add an atomic item (bounded title, bounded description, code locations, gates — all required).

Read one item with full untruncated detail.

Update item metadata: note, title, description, high_priority, and/or replacement dependency ids.

Resolve an item as completed with commit id + gate results.

Resolve an item as split, creating a seeded subsprint atomically.

Resolve an item as deprecated with a reason.

Add a note to a specific item id. Notes are not substitutes for item_add(); create items for trackable work.

Read notes attached to an item id.

List notes for an item as compact rows.

Update a note by note id.

Attach a file path to the sprint, optionally related to item ids.

Read one artifact by id.

List active sprint artifacts as compact rows.

Update an artifact by id.

Regex search over sprint text with character context around each match.

Render the sprint changelog as Markdown with change-map and coverage tables.

Return the current dashboard URL and port without starting or restarting it.

Restart the follow-along dashboard and return its URL and port.