# Scraper triage checklist

A six-section checklist for diagnosing a broken scraper in order:
selector drift, empty fields, pagination loops, rate limits, login
walls, and output checks. Each section is a short list of concrete
checks with the fix pattern next to the symptom.

## Files

- CHECKLIST.md: the checklist itself. Work top to bottom and stop
  at the first section that explains the symptom.
- sample-triage-report.md: a filled-in triage using a real defect
  from the Hacker News front-page scraper, verified 2026-11-12.

## Use

Copy CHECKLIST.md into the failing project, work through it
against the live page, and write the outcome into a triage report
shaped like the sample: symptom, checklist path, fix, and how the
fix was verified. Keep the report with the project so the next
break starts from evidence.

There is no code in this artefact and nothing to install.

## Verification log

- 2026-11-19: checklist and sample report written and read back.
- The worked example states only verified facts from the HN
  pipeline: the `.hnuser` class is gone from live Hacker News, the
  submitter column came back empty while mocked tests passed, the
  fix uses `a[href^="user?id="]`, and the live rerun on 2026-11-12
  produced 30 rows with submitter populated and 6 of 6 tests
  passing.
