Review Judge — Development agent for Claude Code
Neutral zero-context judge for the adversarial-review skill.
How to install Review Judge
Installs to ~/.claude/agents/ksdisch-claude-config-review-judge.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/ksdisch/claude-config/HEAD/agents/review-judge.md -o ~/.claude/agents/ksdisch-claude-config-review-judge.md Restart Claude Code, or start a new session, for it to be picked up.
What Review Judge does
name: review-judge description: Neutral zero-context judge for the adversarial-review skill. Rules ONLY on findings the author disputed — reads the reviewer's claim, the author's rebuttal, and the actual code at HEAD, then rules UPHELD / OVERRULED / DOWNGRADED / UPGRADED with one paragraph of reasoning appended to the review mailbox. Owes deference to neither side and never edits code. Do NOT auto-delegate or launch proactively; use only when the adversarial-review skill (or Kyle, explicitly
Alternatives in Development
- Get Commit ID — gh api repos/OWNER/REPO/pulls/PR_NUMBER --jq '.head.sha' gh api repos/OWNER/REPO/pulls/PR_NUMBER/comments \ 1.3k ★
- Finding Checker — Blind adversarial checker — given ONLY a finding artifact and its evidence (never the author's reasoning), tri 348 ★
- Cold Verifier — Phase 11 cold verification agent that independently re-verifies CRITICAL and HIGH findings with zero prior con 125 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Debate Tribunal
Multi-agent structured debate protocol where 3-5 persona-driven debaters argue positions through claim-evidenc
Psychodrama Judge
Use ONLY by psychodrama-protocol skill orchestrator. Aggregates R1 panel votes into a structured verdict; iden
Acceptance Phase
Enforcement playbook for the fresh-context UAT acceptance gate: capability preflight, applicability/risk class
Changelog Updater
Parses commits between the most recent tag and HEAD, then appends Keep a Changelog-formatted entries to CHANGE
Checklist Verifier
Verifies a single claim from the verification checklist against the actual source code. Reports PASS, FAIL, or
Completion Gate
Lie detector for agent claims — validates every completion claim has actual evidence. Default-FAIL mindset. Us
Related Skills
Reviewer Application
You are reviewing one focused group of changed application files from a pull request for an automated pipeline
Claim Review
Run the three reviewer agents in parallel over a scope and consolidate their findings into one report.
Reviewer2
Invoke Reviewer 2 adversarial review on findings