The knowledge corpus (CORRAL.md)

A repo that runs with corralai can carry its working knowledge as a markdown corpus in the repo itself: CORRAL.md at the root as the entry point, docs/corral/*.md as the corpus. The same corpus serves four readers — developers read it as onboarding docs, any developer’s coding agent queries it conversationally, the herd itself searches it before working and extends memory as it learns, and it grows the way code does — through ordinary pull requests, where code review is the same gate that vetted memory passes through.

Advisory, not injected

The corpus is ingested as shared=false, tagged to its repo — it’s found by search_memory, the same way a herd-written lesson is, but it never auto-injects into phase instructions the way vetted, human-promoted guidance does. Repo-shipped knowledge earns its trust through code review at write time; herd-proposed knowledge earns it through the operator’s approval click at learn time — two different gates for two different origins of guidance, landing in the same searchable place.

The loop closes here

Skills the herd proposes and a human approves land in the same searchable memory as this corpus — herd-discovered knowledge and developer-written knowledge accumulate side by side, queryable by every agent that joins.

This site’s own docs are built the same way this page describes: sourced from README.md, docs/DESIGN.md, and docs/corral/*.md, restructured for a sidebar rather than duplicated by hand.