Study Analysis and Reporting
Dual-agent study analysis with cross-verification.
Overview
Two agents analyze your data in parallel and independently, then a third pass cross-checks them and flags where they disagree. You get a written summary plus annotated R or Python code you can re-run yourself. Before any of that, a short interview checks whether your documentation is complete (codebook, preregistration, analysis template) and adapts: if everything is in place it confirms in one pass, otherwise it asks targeted questions about variables, conditions, and analyses. Bundles study-summary (output templates and AMA formatting) and methodological-detail-log (decision audit trail).
When to use
- To automate the analysis and internal reporting for a new study
- To get annotated R or Python code you can re-run yourself
- To generate reproducible code and a structured statistical report for old studies
- To document every decision that goes into an analysis
- To efficiently ingest a new dataset into an existing project folder
What’s in the box
This download installs 3 skills that work together. Unzip and the three folders sit side-by-side in your skills directory.
verified-study-analysis— The dual-agent orchestrator with the pre-analysis interviewstudy-summary— Analysis code and .docx summary templates with AMA formattingmethodological-detail-log— Logs analytical decisions for the audit trail
Download
Unzip into ~/.claude/skills/ (Claude Code) or ~/.codex/skills/ (Codex), then restart your CLI.
Feedback
Found a bug or have a suggestion? Email me.