What problem does it solve? Running a group of planned stories one by one in a fresh session risks losing progress mid-batch, leaving stories half-flipped between statuses, and exhausting context before the batch finishes. This Skill runs one chat (batch) from an Implementation Strategy end to end with atomic per-story finalisation, a context-budget guard, and a clean failure-recovery contract. ## Core Features & Use Cases - Sequential story loop: Composes execute-story, run-testplan, and close-out-story for each story in the chat, respecting lane ordering and DoR/DoD gates. - Atomic finalisation and recovery: Each story commits and flips to done before the next starts; on abort, completed stories stay done, the in-flight story is marked blocked, and remaining stories stay ready. - Context-budget guard: Aborts cleanly before starting a story that would push context utilisation above roughly 80 percent, so the batch can resume in a new session. - Use Case: After running /tandem:execution-strategist, invoke /tandem:execute-batch CHAT-01 to implement, test, and close every ready story in that chat, then flip the chat's executed flag and regenerate the dashboard. ## Quick Start Ask the assistant to run /tandem:execute-batch with a chat id such as CHAT-01 from the latest execution strategy.
