What problem does it solve? Web fetches frequently fail with 403/429 errors, paywalls, Cloudflare interstitials, or bot-detection walls, leaving you without the content you need. This Skill provides an ordered fallback ladder that retrieves third-party copies of the page instead of retrying the same blocked URL. ## Core Features & Use Cases - Multi-route recovery ladder: Tries the Wayback Machine, archive.today domain rotation, and Jina Reader (when JINA_API_KEY is set) in cheapest-first order via the bundled recover_page.py script. - Fake-success detection: Validates every candidate body with byte floors, interstitial-title rejection, and redirect-stub detection so dead Google Cache pages, AMP stubs, and 429 rate-limit bodies are never mistaken for real content. - Provenance discipline: Labels every result as a dated snapshot or a live re-render so citations never present stale archives as current pages. - Use Case: A news article returns a paywall interstitial. Run the script with the URL and receive the first genuine archived copy along with its snapshot timestamp and source URL, ready to cite as "as archived on <date>". ## Quick Start Ask the agent to recover the blocked article at a given URL using the blocked-page-recovery script and report its provenance.
