SolidMITOriginal

Dead Link Crawler

Crawls a live site using a Python stdlib crawler and reports internal links returning 4xx/5xx status codes.

Entry verified April 21, 2026

The short answer

Point it at one site name, or run against all Netlify-hosted sites with no argument. It walks up to 40 internal pages per site, follows same-domain anchors, and groups broken links by site so you can triage 404s fast.

When to use it

Run after a URL-structure change or a site restructure (e.g. Vendors.ie or BAWT) when internal links are most likely to snap. Also useful as a pre-audit sweep before a full QA pass, or when outbound references have had time to rot and need a spot-check.

Setup

  1. 1

    Save the file as ~/.claude/commands/dead-links.md.

  2. 2

    Invoke with an optional site name or number from the /sites list: /dead-links <site>.

  3. 3

    With no argument, the crawler runs against all sites in the built-in list (vendors.ie, tenderwatch.ie, fluentatdesk.com, the flash reports, readingcurator.com, glitchtexteffect.com, mytriathloncalculator.com).

  4. 4

    Uses urllib only — no external dependencies, no wget required.

  5. 5

    For deeper crawls, run against a single site since the 40-page cap keeps batch runs reasonable.

Example

You: /dead-links promptwritingstudio.com
Claude: Crawled 184 pages, 1,247 links. 3 broken. All 3 are outbound (rotted referenced articles). Fix list attached.

Source & attribution

Author
Bryan Collins
Licence
MIT (author's own work)
Type
Original

Original pattern published under MIT — attribution preserved by convention, not licence requirement.

Caveats

Respect robots.txt on outbound crawls. Heavy on big sites — rate-limit.

New skills, explained plainly.

One short email when a new Claude Code skill is worth installing — and when to skip the hype. No spam.

56 skills across 10 categories, all licence-verified.