# Full catalog & upstream

This skill is a self-contained methodology. The complete, continuously-maintained catalog it draws on — **29 search channels, 460+ statistical sources, 39 validated APIs, and 103 report blocks** — lives in the upstream project, where a weekly CI job re-validates every API endpoint and flags dead links:

**https://github.com/Socialpranker/claude-deep-research**

What's upstream (kept there to avoid duplicating a large, fast-moving catalog):

- `phases.yaml` — the canonical phase definitions and depth gating.
- `references/channels.md` — the 29 named search channels.
- `references/stat_sources/` — the 460+ statistical-source registry.
- `references/api_sources/` — the 39 API definitions with auth and fallback notes.
- `references/blocks/` — the 103 report building blocks per genre.

The `SKILL.md` here is the operating discipline; pull the upstream catalog when you want the full sourcing surface or the auto-validated API list.