What problem does it solve? Ending a coding session cleanly requires several easy-to-forget steps: confirming all work is committed, archiving finished tasks, and writing a session journal. This Skill automates that wrap-up sequence so nothing is left uncommitted or undocumented. ## Core Features & Use Cases - Dirty-tree classification: Inspects git status --porcelain, filters out Trellis-managed paths, and decides whether remaining changes belong to the current task or parallel work, blocking the wrap-up if uncommitted task code exists. - Task archiving: Archives the active task (plus any user-confirmed completed tasks) via task.py archive, producing automatic chore(task): archive commits. - Session journaling: Records a session entry with title, commit hashes, and summary through add_session.py, producing a chore: record journal commit. - Use Case: After finishing a feature and committing code in workflow Phase 3.4, run this Skill to archive the task and log the session journal in the correct commit order. ## Quick Start Ask the AI to finish the current Trellis work session by archiving the active task and recording the session journal.
