playwright-evidencelisted

# Playwright Evidence Skill ## Overview This skill captures evidence and stores it under `.copilot-tracking/skraft-plans/{projectSlug}/changes/{date}/evidence/{story}/evidence/`, keeping the story key consistent with all other SDLC phase artefacts. Agents read the manifest and decide what to publish. ``` E2E Test Run → Capture Evidence → Write Manifest → Agent publishes Playwright screenshots .copilot-tracking/skraft-plans/{projectSlug}/ (orchestrator (on failure) videos deliver/{story}/ or other agent) traces evidence/ report manifest.md ``` `{story}` is the story slug passed as `SKRAFT_STORY_ID` env var (e.g. `42-add-eligibility-check`). All other phases follow the same convention: `discuss/ac-draft-{story}.md`, `design/diagrams-{story}.md`, `distill/impl-plan-{story}.md` — evidence is no exception. **Scope of this skill:** capture, name, store, and list evidence. Publishing is the agent's responsibility. ## Docker Dependency Freshness (MANDATORY) Before running ANY e2e verification — first run or re-run after a fix — rebuild and restart every Docker Compose / Testcontainers dependency the suite talks to (`docker compose down && docker compose up -d --build`, or the equivalent Testcontainers container recreation). Do NOT reuse a container left running from a previous session or a previous story. **Why:** a stale container silently passes tes