Experiment Runner — Data & AI agent for Claude Code
ML/empirical experiment design specialist — pre-registration, Fisher-style design, reproducibility manifests.
How to install Experiment Runner
Installs to ~/.claude/agents/cdeust-zetetic-team-subagents-experiment-runner.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/cdeust/zetetic-team-subagents/HEAD/agents/experiment-runner.md -o ~/.claude/agents/cdeust-zetetic-team-subagents-experiment-runner.md Restart Claude Code, or start a new session, for it to be picked up.
What Experiment Runner does
name: experiment-runner description: "ML/empirical experiment design specialist — pre-registration, Fisher-style design, reproducibility manifests" model: sonnet effort: medium when_to_use: "When an experiment is about to be designed, run, or reported." agent_topic: experiment-runner tools: [Read, Edit, Write, Bash, Glob, Grep, mcp__plugin_hypermnesia-mcp_cortex__unified_search, mcp__plugin_hypermnesia-mcp_cortex__recall, mcp__plugin_hypermnesia-mcp_cortex__remember, mcp__plugin_hypermnesia-
Alternatives in Data & AI
- Credential Hunter — Autonomous credential-attack pipeline runner 4.5k ★
- Database Architect — Database architect for data modeling, technology selection, schema design, and migration planning 2.8k ★
- Style Distiller — 三阶段 13 模板 LLM 提取,写风格主卡、场景卡、分析稿与版本快照 669 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Research Validator
Validates scientific correctness, reproducibility, and statistical rigor of research code and experiment desig
Pan Experiment Runner
Drives an external AI coding session against an experiment folder. Observation-only — read-only relative to PA
Foley
Backlog triage analyst — stage 1 of the backlog pipeline. Invoke with pre-fetched task data to assess dispatch
Omn Dev 1 Implement.Agent
Implementation Developer of the AI Engineering Framework. Implements approved changes with production quality
Add Task Type
Adds a new AI task type to the woolball server: endpoint, handler, preprocessing route, logic class, DI regist
Mle Reviewer
Production machine-learning engineering reviewer for data contracts, feature pipelines, training reproducibili
Related Skills
Experiment Design
Design empirical studies through power analysis, pre-analysis planning,
Flow Experiment
Experiment flow backed by manifests, experiment summary state, and shared middleware
Counterfactual Statistics
Statistics for counterfactual and paired-branch ML experiments - the independent unit, paired tests against a