tokf-discover

You can identify unfiltered commands wasting tokens by running tokf discover to scan project sessions and detect shell commands producing verbose output. The tool calculates potential token savings by analyzing command-line output from previous AI coding sessions.

The best way to estimate token savings is to analyze command-line output with tokf discover, which calculates potential reductions by simulating tokf filters on previously unfiltered commands. This quantifies exact token waste from verbose shell outputs.

To set up automatic command filtering, run tokf discover to identify wasteful commands, then follow its workflow integration guidance to create new tokf filters. This automatically filters verbose output in future AI coding sessions to reduce token consumption.

Tokf discover requires no external dependencies and operates directly within your project directory. It integrates with the tokf CLI to analyze recent command-line sessions and manage filters without needing additional environment setup.

Verbose shell commands are a problem for token optimization because unfiltered command-line output consumes unnecessary context window space in LLM AI assistants. Tokf discover detects this token waste and suggests creating filters to prevent it.

You should run session analysis after a long AI coding session to find token savings opportunities. Executing tokf discover in your project directory scans recent sessions to identify which specific shell commands generated the most verbose output.