km ai

Connects Kimün with coding agents and language models. It has three subcommands: skill and permissions set up Claude Code to use km, and analyze asks a model to write a report on its own.

km ai skill claude
km ai permissions claude
km ai analyze claude [path]

Every provider argument is claude, the only one supported today.

km ai skill — Install the skill

Installs a Claude Code skill that teaches the agent how to run the km subcommands and how to read their JSON output, so it measures before it changes anything.

km ai skill claude                      # install the skill
km ai skill claude --with-permissions   # and let km run without prompting

No API key is needed: Claude Code itself is the model.

Flag	Description
--with-permissions	Also configure the permissions, as km ai permissions does

km ai permissions — Run without prompting

Adds Bash permission rules for every km subcommand to the .claude/settings.local.json file of the project, so Claude Code runs them without asking each time. It merges with the permissions already there and is safe to run more than once.

km ai permissions claude

km ai analyze — A report written by a model

Invokes a model that uses the km tools to analyze the repository and writes a report: code health, complexity hotspots, maintainability problems and recommendations.

km ai analyze claude                      # analyze the current directory
km ai analyze claude src/                 # analyze a subdirectory
km ai analyze claude --output report.md   # save the report to a file

It calls the Anthropic API, so it needs the ANTHROPIC_API_KEY environment variable.

Flag	Description
--model MODEL	Model to use
-o, --output FILE	Save the report to a file