Review Arbiter — Data & AI agent for Claude Code
Authoritative LLM-as-a-Judge meta-agent that adjudicates multi-agent plan and code reviews whenever two or more reviewers disagree, filtering false positives, hallucinations, and out-of-scope nitpicks.
How to install Review Arbiter
Installs to ~/.claude/agents/ivanestar-eitr-review-arbiter.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/ivanestar/eitr/HEAD/.claude/agents/review-arbiter.md -o ~/.claude/agents/ivanestar-eitr-review-arbiter.md Restart Claude Code, or start a new session, for it to be picked up.
What Review Arbiter does
name: review-arbiter description: Authoritative LLM-as-a-Judge meta-agent that adjudicates multi-agent plan and code reviews whenever two or more reviewers disagree, filtering false positives, hallucinations, and out-of-scope nitpicks before issuing the actionable Arbiter Verdict. tools: Read, Grep, Glob model: sonnet
Role: Review Arbiter (The Review Judge)
You are the authoritative Review Arbiter and Quality Judge for all multi-agent plan and code reviews in this repository, invoked
Alternatives in Data & AI
- Svelte Correctness Review — Reviews a feature segment in any SvelteKit app (apps/moderator, apps/auth, apps/creator-studio) for correctnes 7.2k ★
- Visual Explainer — Agent skill + prompt templates that generate rich HTML pages for visual diff reviews, architecture overviews 6.8k ★
- OSA Operations — Meta System — You are inside the OSA Operations repository 174 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Classifier Tuner
Improves the "data tender" keyword classification of gebiz-radar. Samples classified and unclassified tenders,
Council Chairman
Chief arbiter of the LLM Council that synthesizes multi-model responses. MUST BE INVOKED after council members
Judge Writer
Writes a binary LLM-as-judge prompt for exactly one failure mode, from a definition and a handful of human-lab
Refactor Prompts Judge
LLM judge for /refactor-prompts. Classifies hot regions from the lexical scan and emits an edit plan JSON.
Metis
Pre-planning gap analyzer. Catches hidden intentions, ambiguities, and AI failure points BEFORE plan generatio
AI Implementer
Implements ONE approved plan step in a project with .ai/ — mechanical, pattern-copying work where the step nam
Related Skills
Makoto
When Claude says “tests pass”, makoto checks the record — every claim held against the agent's own logged deed
Advanced Code Review Verify
Advanced Code Review Phase 4: Verification - fact-check findings against codebase, remove false positives
Klarion
AI-powered secret scanner for code and git history. Finds leaked API keys, tokens and credentials with near-ze