GitHub · TypeScript
natea/harness-eval
Eval framework that ranks agentic coding frameworks (Superpowers, Compound Engineering, Agent Skills, GSD) by having each build the same spec with the same harness+model in isolated sandboxes, graded on a weighted evidence-based rubric: PRD adherence, code quality, speed, token spend
- Stars
- 15
- Forks
- 5
- Stars gained in 24 hours
- 0
- Forks gained in 24 hours
- 0
Topics: ai-agents, benchmark, bun, claude-code, coding-agents, eval, llm, vibench
Latest data point: Nov 21, 2026 · 15 stars · 5 forks