Skip to content

Claude Skills by closestfriend

github.com/closestfriend

1 skillA× 10 installs0 views

benchmark-designerADesign, red-team, and ship rigorous evaluation suites, benchmarks, and eval harnesses for AI models and agents. Use whenever the user wants to measure, compare, or score model or agent performance — building an eval, benchmark, test harness, scorer, grader, verifier, rubric, LLM-as-judge, leaderboard, regression suite, or task set — including when they only describe the goal ("I want to see which model is best at X", "how do I test whether this agent can Y", "set up evals for my MCP server", ...ai-agentspythonrustVotes: 0GitHub stars: 2