Chief Evaluator — Research agent for Claude Code
Use for strategic evaluation decisions, tier selection, methodology oversight, and cross-domain coordination.
How to install Chief Evaluator
Installs to ~/.claude/agents/homericintelligence-scylla-chief-evaluator.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/HomericIntelligence/Scylla/HEAD/.claude/agents/chief-evaluator.md -o ~/.claude/agents/homericintelligence-scylla-chief-evaluator.md Restart Claude Code, or start a new session, for it to be picked up.
What Chief Evaluator does
name: chief-evaluator description: Use for strategic evaluation decisions, tier selection, methodology oversight, and cross-domain coordination. Invoked for high-level evaluation strategy and research direction. tools: Read,Write,Edit,Bash,Grep,Glob,Task,WebSearch model: opus
Chief Evaluator Agent
Role
Level 0 Meta-Orchestrator responsible for strategic evaluation decisions across all of Scylla. Sets evaluation strategy, selects tiers for comparison, and approves methodology.
Alternatives in Research
- Peer Reviewer Qual — A simulated peer reviewer specializing in qualitative and ethnographic methods — in-depth interviews, ethnogra 134 ★
- Ds Pro — DSH (DeepSeek Harness) worker on the pro tier - stronger reasoning 116 ★
- Finding Extractor — Extracts atomic, testable findings from a single research paper 95 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Cpo
Chief Product Officer who generates innovative business ideas with market analysis, competitive research, and
Statistical Specialist
Use for statistical analysis, hypothesis testing, and significance testing. Invoked for comparing tier results
Rc Evaluator
研究サイクルの「分析・考察」フェーズを担う.実験結果がノイズか有意な変化かを判定して解釈し,採否を判定して journal を確定,次に振るレバーを決めてイテレーションを git commit する.research-c
Guide Writer
(/auto, /meeting) Maintain research//guide.md as the human oversight entrypoint, reading durable node surfaces
Execution Specialist
Trade execution specialist for order management, execution algorithm selection, transaction cost analysis, and
Robium Architect
Robotics application architect. Use PROACTIVELY at the start of any new robotics application to run the heavy
Related Skills
Agentic Eval
Niche-agnostic agentic evaluator using CLEAR v2.0 framework — 6-domain assessment, 8 analysis dimensions, 6-ti
Branding Harness
Claude Code skill: generate an integrated brand identity system (strategy + verbal + visual direction) with a
Council Gate
Council direction gate (per fase, alleen na een evaluator-ACCEPT)