capturing-story-contextlisted

# Capturing Story Context Produce a confirmed context block in `.claude/qa/<story-id>.md`. Generation does not start until this exists. ## 1. Get the identifiers **The work item ID is all you need.** IDs are unique across the organisation, so the project is not required to fetch a story — it is read from the story itself and used later for publishing. QA usually pastes a URL — from the board, a backlog, a sprint taskboard, a query result, or the item itself. Do not try to recognise each shape. Take the number: - the `workitem=` query parameter when present, - otherwise the numeric segment following `_workitems/edit/` (a trailing slash is not part of the ID). ``` …/_workitems/edit/24073/ -> 24073 …/_boards/board/t/apptor-webhook%20Team/Stories?workitem=24073 -> 24073 …/_backlogs/backlog/apptor-webhook%20Team/Stories?workitem=23471 -> 23471 ``` A bare ID is equally fine. The organisation comes from `az` defaults, or from the URL host when working across organisations. Never search for the story, never list projects, never fetch unrelated items. The scope of a run is exactly one story. **Echo what you resolved** — `24073 — Provider self-signup UX (apptor-webhook)` — before doing anything else. This is what makes an unrecognised URL shape harmless: either QA sees the right story named back, or they correct you immediately. It costs one line and removes the need to anticipate every URL form. ## 2. Fetch the story `fetch_story