gather-documentation

To gather authoritative documentation without fabrication, use a discovery, retrieval, and synthesis pipeline that checks for llms-full.txt at official docs sites and audits every technical claim against fetched sources. This ensures specific API names, payload fields, and error codes are traced to verified material.

An llms-full.txt file is an LLM-consumable documentation bundle provided by some official docs sites. You need it when building skills or researching technologies, as it serves as complete ground truth for ingesting and synthesizing structured reference material without missing context.

To retrieve official docs and GitHub examples for creating a structured skill outline, first ingest llms-full.txt or llms.txt, then retrieve key official pages and supporting GitHub repositories, and finally synthesize the fetched content into skill-ready material covering APIs, configuration, and gotchas.

Yes, you can use community GitHub context alongside official documentation. A synthesis pipeline identifies official docs sources first, then adds relevant GitHub repository patterns and issue gotchas to provide complementary context for creating comprehensive skill references.

Documentation synthesis requires a fabrication audit to prevent invented details. It verifies that every specific technical claim, such as API defaults, payload fields, and error codes, is explicitly supported by retrieved sources, ensuring provenance and accuracy in the final structured skill output.

The best way to find LLM-consumable doc bundles for unfamiliar APIs is checking the official docs site for llms-full.txt and llms.txt files. Prioritizing llms-full.txt provides complete ground truth, which can be combined with GitHub examples for structured skill research.