workloglisted

# worklog — the live working-memory comm A **worklog** is a Comment.io comm that is a faithful, current mirror of the material state of one task: topology, plan, status, decisions, receipts, non-technical summary, and open questions. It is shared working memory, not a command transcript. This skill is a **primitive**: `comment-feature` and `comment-bug` build on it. Use it standalone only when a human explicitly asks for a watchable, steerable record — not automatically for any multi-step task. ## When to use - When a human explicitly asks for a watchable/steerable record, or when a delivery skill (`comment-feature` / `comment-bug`) opens one for its task. - Not for routine local, read-only, or investigative work. - Once opened, create the worklog *first*, then work, updating as you go. ## Project Root behavior - If a delivery skill calls `worklog` with no existing Project Root, the worklog is the Project Root. After creation, update `Project Root: URL` near the top with this worklog's human-openable URL. - If the caller passes an existing Project Root, this worklog is a child. Add `Project Root: URL` near the top using the caller's root URL, and link the worklog from the root. - Standalone worklog behavior is unchanged. Do not invent a Project Root unless a full delivery flow or the human asked for one. - Keep the body authoritative and scannable. Put process logs, bulky evidence, FYI detail, and review rounds in comments or linked child docs. ## The comm structure (b