Traction scoreGitHub stars can be faked, so popularity alone can be misleading. Traction Score looks for broader signs of recent attention, adoption, and active maintenance.
Started as a combination of caveman and ponytail, with a focus on shorter agent output. The repository publishes raw benchmark transcripts, including losses, so you can inspect the tradeoffs yourself.
WHO IT'S FOR
AI agent power users
reducing token consumption for AI coding agents
Technical leads / architects
auditing repositories for over-engineering and bloat
Compatible AgentsThe repository documents support for these agents. The skills may also work with other agents that can load SKILL.md files, but they may need some setup or small changes.
Replace <file-path> with an existing source file in your project.
Make that file available to the agent.
Make sure chisle-audit is installed in your agent before trying this task.
Use chisle-audit to audit <file-path> for over-engineered code and bloated comments or docstrings. Keep it read-only and rank high-confidence findings by biggest saving first.
An agent session with access to your project and a current diff to review.
Make sure chisle-review is installed in your agent before trying this task.
Use chisle-review to review my current diff for needless abstractions and speculative scaffolding. Report one finding per line with a simpler replacement. Do not edit files.
An agent session with the chisle-help skill available.
Make sure chisle-help is installed in your agent before trying this task.
Use chisle-help to show the quick reference for activating and stopping Chisle, auditing a file, and reviewing a diff. Include the code efficiency ladder.