You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Add official Porkbun and Cloudflare fetch selectors at the documentation fetch owner (scripts/fetch_all_docs.sh).
Validate the Porkbun and Cloudflare LLM manifests independently from the HTML mirrors so truncated fetches fail closed (scripts/lib/documentation_fetch_sources.sh).
Register canonical citation/provenance mappings for both corpora (DocsSourceRegistry).
Cover named-selector dispatch, completeness checks, publication, and citation restoration in repository-owned tests.
Download both corpora without altering the active sole ingestion writer.
Queue both doc sets for sequential embedding after the current repository backlog.
Acceptance Criteria
Porkbun coverage includes API authentication, account/domain/DNS/DNSSEC/SSL/hosting/email/marketplace/pricing/webhook/sandbox/mock surfaces, task guides, and the official MCP agent setup.
Cloudflare coverage includes all 8,292 sitemap pages, including DNS, cache, R2, Wrangler, MCP, AI Gateway, Workers, and Workers AI.
Both LLM manifests pass minimum-size and required-topic validation.
Both local mirrors publish atomically and map local citations back to their official HTTPS pages.
Each new doc set reaches a nonzero exact local Qdrant postcondition through the sole-writer lane.
Summary
Add the complete official Porkbun API/MCP and Cloudflare developer documentation as citation-preserving JavaChat RAG corpora.
Motivation
The current documentation registry and fetch selector contain no Porkbun or Cloudflare source. Porkbun publishes API v3.15 documentation at https://porkbun.com/api/json/v3/documentation, a complete Markdown reference at https://porkbun.com/llms-full.txt, and an official MCP setup at https://porkbun.com/llms/agent-setup. Cloudflare publishes 8,292 canonical documentation URLs through https://developers.cloudflare.com/sitemap-0.xml and the complete 57,269,704-byte corpus at https://developers.cloudflare.com/llms-full.txt.
Tasks
scripts/fetch_all_docs.sh).scripts/lib/documentation_fetch_sources.sh).DocsSourceRegistry).Acceptance Criteria