edonadei/caliper — GitHub trending stats & insights

Busbar

Bifrost

Omnigraph

Kane CLI

edonadei/caliper

#AI agent

Run your real agent with and without your skills, MCPs, and rules. See which ones actually help, and what they cost in tokens. Supports Claude Code, Codex, Pi, and Hermes.

Visit GitHub

Python

194

1 contributors

MIT License

website

Social mentions

Recent discussions about this repository across the web

A skill that works once isn’t necessarily helping your agent. Caliper is an open-source evaluation tool for builders who want to test agent skills with Claude Code, Codex, Pi, or Hermes. It helps you…

@DanKornas · x.com

Following this, I trimmed of my sloppy skills. It reduced in size by 57%. I tested my agent with that skill, it still worked. Amazing. I ran the same test again a week after. This time, it failed.…

@edonadei_ · x.com

I'm doubtful of all the skills you can find in the wild. Built myself a CLI / Skill to run evaluations based on YAML evaluation with pass@k. pip install caliper-eval Will report the gems I find.

@edonadei_ · x.com

Repository activities

repository's daily and monthly activities across stars, forks, merged PRs, issues, and closed issues