Back to browse
GitHub Repository

Lint, benchmark, and score your AI coding instructions. Stop guessing, start measuring.

4 starsTypeScript

agenteval – static analysis for AI coding instruction file

by lukasm1703·Apr 3, 2026·7 points·0 comments

AI Analysis

●●SolidBig BrainNiche Gem

Finally treats AI instructions like code—with linting, benchmarks, and CI gates.

Strengths
  • Harvest command builds eval tasks from git history automatically
  • Catches dead references, contradictions, and token budget overruns statically
  • Self-contained binary requires no runtime—curl install and go
Weaknesses
  • Emerging category means unclear long-term adoption among teams
  • Benchmark quality depends on harvested task relevance to your codebase
Target Audience

Engineering teams using AI coding assistants

Similar To

Promptfoo · LangSmith · Braintrust

Similar Projects