Supervisor Evidence — Development agent for Claude Code
Evidence Judge.
How to install Supervisor Evidence
Installs to ~/.claude/agents/ma-nucho-pro-supervisor-claude-plugin-supervisor-evidence.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/ma-nucho-pro/supervisor-claude-plugin/HEAD/agents/supervisor-evidence.md -o ~/.claude/agents/ma-nucho-pro-supervisor-claude-plugin-supervisor-evidence.md Restart Claude Code, or start a new session, for it to be picked up.
What Supervisor Evidence does
name: supervisor-evidence description: Evidence Judge. Use only when the Supervisor coordinator dispatches this independent role. model: inherit effort: medium maxTurns: 20 disallowedTools: Write, Edit
Evidence Judge
Mission
Audit material claims against supplied files, tool results, tests, and current authoritative web sources. Detect stale evidence, citation mismatch, unsupported inference, source-quality problems, and overclaiming.
Non-negotiable operating rules
You are a
Alternatives in Development
- Mathodology Award Judge — Use for one independent blind judge seat scoring an award-level submission against contest rubric criteria 137 ★
- Chamber Synthesizer — Phase 10 Review Chamber coordinator and judge that orchestrates the debate lifecycle between Ideator, Tracer 125 ★
- Supervisor Agent — You are the supervisor agent for Correctless Auto Mode 51 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Supervisor Architecture
Architecture Judge. Use only when the Supervisor coordinator dispatches this independent role.
Supervisor Completeness
Completeness Judge. Use only when the Supervisor coordinator dispatches this independent role.
Ww Judge 15 Independent Blind Reviewer
Wonder Woman 15 independent blind reviewer. Use only when the Wonder Woman coordinator dispatches this special
Ww Judge 03 Evidence Prosecutor
Wonder Woman 03 evidence prosecutor. Use only when the Wonder Woman coordinator dispatches this specialist ver
Implement Coordinator
Coordinate parallel execution of independent plan tasks. For single tasks — implements directly with quality s
Ranking
Council Ranking role. Judges one blinded pair of hypotheses (A vs B) against the evidence on file and returns
Related Skills
AI Debate Club
A multi-agent debate engine. Two different LLMs argue any motion, research the web for themselves, and cite th
Anti Subagent
FUCK SUBAGENTS. Multi-agent architecture built with autonomous full agents instead of native subagents. Superv
Harness Supervisor
A supervision process several repositories share: one canonical skill for Codex and Claude Code, role routing,