eval-judge
LLM judge for plugin quality assessment. Scores skills on triggering accuracy, orchestration fitness, output quality, and scope calibration using anchored rubrics.
- Score
- Unranked
- Evidence
- No signals yet
- Last commit
- Not known
- Listed
Install
armory install eval-judge --cli claudewrites the file to.claude/agents/eval-judge.mdListed as compatible
# fetches the source and writes it to:
.claude/agents/eval-judge.mdNeeds the armory CLI · not on npm yet, build it from cli/ in the repository
What it is
wshobson/agents sub-agent eval-judge (model: sonnet) from the plugin-eval plugin. LLM judge for plugin quality assessment. Scores skills on triggering accuracy, orchestration fitness, output quality, and scope calibration using anchored rubrics.
When to use it
LLM judge for plugin quality assessment. Scores skills on triggering accuracy, orchestration fitness, output quality, and scope calibration using anchored rubrics.
How to install / invoke
# copy the agent definition into your project's .claude/agents/
curl -sL https://github.com/wshobson/agents/raw/main/plugins/plugin-eval/agents/eval-judge.md -o .claude/agents/eval-judge.md
Notes
Extracted from wshobson/agents. Plugin: plugin-eval.