Skip to content

// Build, run, and report on LLM evals. Pairwise comparisons, judges, regression detection.

Evals RunnerVerified Creator

git log --oneline --stat
HEAD
Stars
2.9k
Forks
98
Updated
Apr 27, 2026
repo --stat

  • stars

    2.9k

  • forks

    98

  • last update

    Apr 27, 2026

  • license

    MITv2.0.5

quickstart.sh
3 steps
  1. Install

    // Drops SKILL.md into ~/.claude/skills/

    $ claude skills add evals-runner
  2. Invoke

    // Run from any project directory

    $ claude --skill evals-runner "help me ship this"
  3. Iterate

    // Re-run with edits — Claude keeps the skill loaded

    $ claude --skill evals-runner "now refactor it"
evals-runner/
references
SKILL.md
readonly
name:
Evals Runner
slug:
evals-runner
version:
v2.0.5
license:
MIT
author:
@evalforge
repository:
github.com/evalforge/evals-runner
categories:
Prompt EngineerQA / Testing
tags:
#evals#llm#benchmarks#judges#regression
description:

Build, run, and report on LLM evals. Pairwise comparisons, judges, regression detection.

features.md
3 capabilities

// What you can do with it

  • Automates the tedious parts of the workflow.
  • Gives Claude the right context, tools, and guardrails.
  • Produces consistent, reviewable output every time.

README.md

evals-runner/README.md
5 sections
Loading README…
@evalforge/index.json
more by author

authored by

@evalforge
view profile

$ cat reviews/

Reviews

// No reviews yet. Be the first.
Loading review form…

$ ls related/

explore all →