Skip to content

Use it

This section is for the kac user: whoever installs the tool, runs it over a corpus, wires it into CI, and builds and installs the plugin. A corpus is one repository of knowledge records kept in git. The plugin is that corpus packaged for Claude Code.

The pages here

  • Getting started installs the tool and runs it against a corpus.
  • One corpus or several helps you decide how many corpora to keep, and how they share what they know.
  • Running it in CI wires the commands that check a corpus into a pull request, on GitHub Actions or Azure Pipelines, and says what those checks ask.
  • Installing the plugin installs a corpus plugin in Claude Code, and turns on the auto-update that brings each new version to its readers.
  • Troubleshooting is what kac prints when something is wrong, and what to do about it.
  • The corpus descriptor covers .corpus.yaml, the one file a corpus writes for itself.
  • Discovery says which files kac opens, which of those count as records, and which it reads for something narrower.
  • Layers says which files in a corpus belong to the framework, which belong to the corpus, and what happens to each when a newer framework arrives.
  • Imports says how one corpus cites records another corpus published.
  • Skills says which skills kac ships, where each one lives, and who can read it.
  • CLI reference gives a page to each command, saying what it does, what it refuses and what it leaves alone.

Where to go next

Getting started is the first page to read if you have not installed kac.