Brainswarm Judge — Development agent for Claude Code
brainswarm judge role.
How to install Brainswarm Judge
Installs to ~/.claude/agents/kyleyhw-agent-brainswarm-brainswarm-judge.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/kyleyhw/agent-brainswarm/HEAD/.claude/agents/brainswarm-judge.md -o ~/.claude/agents/kyleyhw-agent-brainswarm-brainswarm-judge.md Restart Claude Code, or start a new session, for it to be picked up.
What Brainswarm Judge does
name: brainswarm-judge description: brainswarm judge role. Compares anonymised finalist pairs against the frozen rubric and rules on disputed critiques (Phase 6). Dispatched only by the /brainswarm referee; never invoke directly. tools: Read, Write
brainswarm judge
**SCAFFOLD — role prompt not written yet.** Tool allowlist is a draft; enforcement is still to be verified (docs/DESIGN.md §16.1).
Compares anonymised finalist pairs against the frozen rubric and rules on disputed cri
Alternatives in Development
- Domain Referee 973 ★
- Thermo Nuclear Code Quality Review — Independent strict maintainability reviewer for a collected diff and changed-file contents 144 ★
- Mathodology Award Judge — Use for one independent blind judge seat scoring an award-level submission against contest rubric criteria 137 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Brainswarm Rubric Auditor
brainswarm rubric-auditor role. Checks the referee's draft rubric against the brief for missing, misdirected,
Brainswarm Critic
brainswarm critic role. Steelmans, then attacks, anonymised idea cards with justified critiques (target quote,
Skill Eval Reporter
Compares repeated paired execution results using blind A/B methodology and generates a skill effectiveness rep
Judgment Pair Reviewer
Rubric-based reviewer for doctrine-corpus judgment pairs. Reads each generated pair against its source ADR Dec
Best Of N Judge
L0.97 Adversarial Verification — compares N parallel implementation attempts. AC-compliance and coverage are d
Disc Reviewer Ambiguity
The ambiguity judge of the stage-1 discovery review — compares two blind readers' builds and reports every rea
Related Skills
Persona Portability Benchmark
One persona, one frozen memory, N models: how much of an agent's character survives a model swap? Harness + bl
Jev Search Rerank Eval
Does a TypeSafe Jev rerank beat embedding search? Graded relevance eval (9,831 pairs, 164 zh/en queries) over
Architecture Config Pass Frozen Authority Directly
Use when: (1) a coordinator projects a typed config into a second stage-only config object, (2) new CLI/config