AIPython★ 235⑂ 26 forksMIT
edonadei/caliper
EVALUATE AGENTS WITH CALIPER
Caliper runs your real agent with skills, MCPs, and rules and without them, showing what actually helps and how many tokens it costs. Supports Claude Code, Codex, Pi, and Hermes via CLI in Python.
// KEY FEATURES
- Launches the agent with skills, MCPs, and rules and without them to compare effect.
- Measures the number of tokens spent for each configuration variant.
- Supports evaluating agents Claude Code, Codex, Pi, and Hermes via a single CLI.
- Outputs a clear report showing which components actually improve performance.
#ai-agents#claude-code#cli#codex#evals#evaluation#hermes#hermes-agent#hermes-skill#pi#reliability#skills
Open on GitHub →New repositories every 30 minutes
REDDYX AI scans GitHub 24/7 and ships the best AI/ML/Web3 projects to Telegram.
Join on Telegram