claude

agent-evals

Designing, running and reporting agent evaluations and leaderboards: held-constant ablations across model, harness and task adaptation, with integrity checks.

1 on GitHubApache-2.0updated 7 days ago
/

Plugin agent-evals from marketplace skills in zeroasterisk/skills. Commands read from the repo's manifest — 6 days ago.

What it installs

Skills (1)

  • eval-ablation-framing

    Frames agent evaluations, benchmarks and leaderboards as held-constant ablations across three factors: Model, Harness, and Task Adaptation (none, pre-baked skill, or self-customization where the agent builds its own tools first). Use when deciding which eval cells to run next, designing a benchmark

View source on GitHub

More from zeroasterisk/skills

All 20 plugins →

ai-pop

claude

Virtual music artist production suite: lore planning, audio synthesis, visual design, web player compilation, and public endpoint deployment.

6 skills

1

adlc

claude

Agentic Development Lifecycle (ADLC) toolkit integration: phase router, rail-guard hooks, slash commands, and prosecutor agents.

4 skills

1

Full-stack UI development workflows, React component isolation with Storybook, CUJ capturing, and product intent refinement.

4 skills

1

Repository management skills for technical writing, README generation, changelog curation, and commit recaps.

4 skills

1

Software engineering discipline, systematic debugging, Elixir best practices, secure web guidelines, and agent skill authoring.

4 skills

1

Tools and quality evaluation rubrics for architecture diagrams, terminal animations, YouTube/social hero thumbnails, and mathematical explainer videos.

4 skills

1

Maintain agent-evals?

This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.

claude & codex plugins20 plugins