Experiment Manager — Research agent for Claude Code
Delegate to own and advance the research tree — choose the best pending node to expand, prune branches by failure type, and record outcomes.
How to install Experiment Manager
Installs to ~/.claude/agents/taewooopark-maglab-experiment-manager.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/TaewoooPark/MagLab/HEAD/agents/experiment-manager.md -o ~/.claude/agents/taewooopark-maglab-experiment-manager.md Restart Claude Code, or start a new session, for it to be picked up.
What Experiment Manager does
name: experiment-manager description: Delegate to own and advance the research tree — choose the best pending node to expand, prune branches by failure type, and record outcomes. Keeps tree-search state separate from the orchestrator. tools: [physics_check, provenance_query] model: sonnet max_turns: 12 context: isolated
You are the experiment-manager subagent of MagLab.
① Single objective
Own the research tree (§5.12): decide which pending node to expand next, prune branches that f
Alternatives in Research
- Attack Tree Construction — Systematic attack path visualization and analysis 31.9k ★
- Claude Code MCP Behavior Analysis — Results from 15 tests against a minimal Cord MCP server 133 ★
- Backward Reasoner — Backward Reasoner — Deep Probe Phase 5 hypothesis generator applying Pre-Mortem Analysis and Abductive Reasoni 125 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Appsec Attack Tree
Use proactively to generate a formal AND/OR attack tree for the most critical threats in a repository. Produce
Performance Analysis
perf record -g ./program perf report --stdio valgrind --tool=cachegrind ./program cg_annotate cachegrind.out.
Refactor Cleaner
Dead code cleanup and consolidation specialist. Use PROACTIVELY for removing unused code, duplicates, and refa
Gads Keywords
Keyword research analyst. Uses KeywordPlanIdeaService to expand seed terms with monthly search volume, competi
Memory Analysis Agent
Analyses collected session feedback, detects recurring patterns, and proposes concrete skill improvements. Use
Platform Scout
Owns targets/platforms.yaml and research/job-platforms-2026.md: reads each platform's terms, picks the allowed
Related Skills
Unit Test Expand
Increase test coverage by targeting untested branches and edge cases
Analyzing Systemic Governance Loops 20260822
Diagnosing functional failure by interrogating underlying structural mechanics rather than observed behavior o
Cogenity
Cogenity by Kenneth Lynne is a vendor-neutral multi-account manager for Claude Code and Codex. Keep project ac