parse-playwright-summary

Normalize a Playwright JSON report into the canonical functional contract shape.

Type: skill (deterministic, no model call — thin wrapper over parse.mjs) · Used by: func-run-one, corroborate-2-sources · Status: LIVE (2026-07-08)

Prompt

You will normalize a raw Playwright JSON report. There is no judgment here — run the script:

node .github/skills/parse-playwright-summary/parse.mjs reports/<run_id>/results.json <run_id> reports/<run_id>/contract.json

• Exit 0 → contract written; report the script's OK — … line.
• Exit 1 → the run is FAILED. Report the script's FAILED — <reason> line and STOP. Fails-closed cases: missing/invalid JSON, missing stats (not a Playwright report), zero tests found.

Contract shape (what the script writes)

{
  "run_id": "<string>", "suite": "functional",
  "started_at": "<ISO or empty>", "duration_s": <number>,
  "tests_total": <n>, "passed": <n>, "failed": <n>, "skipped": <n>,
  "failures": [ { "test": "<suite > spec title>", "file": "<spec file>", "error": "<first line>" } ]
}

flaky results count as failed (retries are 0 by convention — flake detection is corroborate-2-sources' job, never Playwright retries).

Tools

Bash (node), Read — no MCP, no model calls.

Data

• Reads: reports/<run_id>/results.json
• Writes: reports/<run_id>/contract.json

Hard rules

1. Never hand-parse the report — the script is the parser; its fixtures live in tests/fixtures/functional/.
2. Fails closed — a malformed report is a FAILED run, never an empty-green contract.
3. Behavior check: powershell -File tests/func-contract-check.ps1.