claude

llm-harness

Builds an evaluation harness around any LLM-dependent system: infers intent from its prompts/code, derives gradeable success criteria, generates golden + adversarial + holdout cases, wires the cheapest adequate grader, and reports a baseline accuracy number to iterate against.

0 on GitHubupdated 15 days agoagent-infra
/

Plugin llm-harness from marketplace skills-personal in dkritarth/skills-personal. Commands read from the repo's manifest — 12 days ago.

What it installs

Skills (1)

  • llm-harness

    Constructs an empirical evaluation suite for LLM components (prompts, tool chains, subagent workflows). Defines measurable assertions, test cases, and scoring metrics to benchmark accuracy.

View source on GitHub dkritarth.com

Problems found in this repo's manifests

  • "mattpocock-skills" points to ./skills/mattpocock-skills, which doesn't exist in the repo — install will fail
  • "research-paper-writing" points to ./skills/research-paper-writing-skills/research-paper-writing, which doesn't exist in the repo — install will fail

More from dkritarth/skills-personal

All 31 plugins →

claude-api

claude

Claude API and SDK documentation skill for building LLM-powered applications

17 skills

0

document-skills

claude

Collection of document processing suite including Excel, Word, PowerPoint, and PDF capabilities

17 skills

0

example-skills

claude

Collection of example skills demonstrating various capabilities including skill creation, MCP building, visual design, algorithmic art, internal communications, web testing, artifact building, Slack GIFs, and theme styling

17 skills

0

agents-sdk

claude

Imported skill: agents-sdk.

1 skill

0

cloudflare

claude

Imported skill: cloudflare.

1 skill

0

Maintain llm-harness?

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 plugins31 plugins