Klein Experimenter — Research agent for Claude Code
EXPERIMENT-loop worker for Klein Auto Research — runs disciplined one-candidate v2 transactions with immutable evidence and track-specific frontiers.
How to install Klein Experimenter
Installs to ~/.claude/agents/xiang-shan-klein-auto-research-klein-experimenter.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/Xiang-Shan/klein-auto-research/HEAD/.claude/agents/klein-experimenter.md -o ~/.claude/agents/xiang-shan-klein-auto-research-klein-experimenter.md Restart Claude Code, or start a new session, for it to be picked up.
What Klein Experimenter does
name: klein-experimenter description: EXPERIMENT-loop worker for Klein Auto Research — runs disciplined one-candidate v2 transactions with immutable evidence and track-specific frontiers. Invoke to "run experiments", "continue the loop", "try the next idea", or "improve the metric" once all gates have cleared. Invoked by /klein run. tools: Read, Grep, Glob, Bash, Write, Edit model: sonnet
FIRST ACTION every iteration: read the study's `playbook.md` (rolling map — current best, ruled-out
Alternatives in Research
- Council Meadows — Council member 4.2k ★
- Worker Researcher — Web research agent 1k ★
- Troubleshoot Agent — Deep research agent for build failures and JUCE problems 209 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Klein Referee
REFEREE worker for Klein Auto Research — Gate 3. Reviews a synthesized study in a FRESH context on a different
Klein Sweeper
SWEEP worker for Klein Auto Research — runs the one sanctioned mechanical parameter search, retaining every tr
Experiment Fixer
Use this agent to diagnose and fix failed AIChor experiments. Reads analysis reports and loop state, explores
Experiment Tracker
PROACTIVELY use this agent when experiments are started, modified, or when results need analysis. This agent s
Study Note Worker
Runs bounded transcript review and source mapping after deterministic Python preprocessing; does not author no
Market Iq
Track competitive landscape, produce competitive moat maps and win/loss analysis. Use when evaluating strategi
Related Skills
Gtm Experiment
Register, track, and analyze A/B experiments
Claude Hook Experiment
Reproducible Claude Code hook experiments for validating settings-based hook behavior
Design Experiments
Draft a locked experiment plan with integrity-hash propagation, parallel research agents, and reviewer subagen