Meta Research

Meta-Research

A Claude Code skill that guides you through a hypothesis-driven research lifecycle — from literature survey to publication — with built-in rigor, reproducibility tracking, and bias mitigation.

What it does

Meta-Research acts as an autonomous research copilot with a **6-phase hypothesis-driven workflow**:

1. Literature Survey — Understand SOTA, identify gaps, open problems, and underexplored areas
2. Hypothesis Generation — Generate broad testable hypotheses, maintain a hypothesis tree in YAML
3. Judgment Gate — Evaluate each hypothesis: novel? important? feasible? falsifiable? already solved?
4. Experiment Design — Rigorous per-hypothesis protocol with locked analysis plans
5. Experiment Execution — Run experiments, track results, determine outcomes
6. Reflection — Analyze results, decide: go deeper, go broader, pivot, or conclude

The workflow is a **research loop** — after experiments, reflection decides whether to continue iterating (generating new hypotheses, running more experiments) or conclude and move to writing.

Two core artifacts track all project state:

• research-tree.yaml — Hypothesis hierarchy with judgments, experiments, and results
• research-log.md — Chronological timeline of exploration and decisions

Installation

From marketplace

/plugins marketplace add 
/plugins install meta-research

Manual installation

# Personal skill (available in all projects)
ln -s /path/to/meta-research ~/.claude/skills/meta-research

# Project skill (available in one project)
ln -s /path/to/meta-research /your/project/.claude/skills/meta-research

Usage

/meta-research [your research question or topic]

The skill always starts with a literature survey to understand the field before generating hypotheses. If you have existing artifacts (`research-tree.yaml`, `research-log.md`), it will resume from the current state.

Examples

/meta-researc