skills.rest
mrrts
mrrtsCOMMUNITY
Explore all skills by this author →
GitHub

run-experiment

Design and execute commit-bound natural experiments on live message corpora.

1|Updated Feb 19, 2026
Search Tags:#data-analysis#rubrics#ai-prompt#experiment-design#natural-language-processing#software-research#worldcli
One-click install
npx skills add https://github.com/mrrts/WorldThreads --skill run-experiment-mrrts
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: run-experiment
Source: https://github.com/mrrts/WorldThreads/tree/main/.agents/skills/run-experiment
Command: npx skills add https://github.com/mrrts/WorldThreads --skill run-experiment-mrrts

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Design and execute rigorous commit-bound natural experiments against a live message corpus to validate hypotheses and measure the impact of prompt-stack changes.

Core Features & Use Cases

  • Structured hypothesis audition with 2–3 candidate hypotheses
  • Pre-registered predictions and explicit success/failure criteria
  • End-to-end experiment design, execution, and reporting with traceability to commits
  • Per-character or per-group scope evaluation and registry integration for reproducibility
  • Documentation-driven workflow that links to reports and project rubrics

Quick Start

Outline a hypothesis, choose a commit boundary, and run a 12-message evaluate with a pre-registered rubric.

Frequently Asked Questions about run-experiment

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run natural experiments on a live message corpus?▼

Natural experiments on a live message corpus are run by defining 2–3 candidate hypotheses, pre-registering a prediction with explicit success criteria, and executing against commit boundaries with a structured rubric for honest interpretation.

What is a commit-bound experiment for prompt-stack validation?▼

A commit-bound experiment validates prompt-stack changes by anchoring natural experiment execution to specific commit boundaries in a live message corpus, ensuring traceability and reproducibility of the measured impact.

How do I pre-register a prediction for data analysis experiments?▼

Pre-registering a prediction involves defining explicit success and failure criteria alongside a structured rubric before execution, then recording the honest interpretation of results in a registered report for full reproducibility.

Can I evaluate prompt changes with a structured rubric on live data?▼

Yes, evaluating prompt changes on live data uses a structured rubric across per-character or per-group scope boundaries, executing a defined message evaluation to measure impact and record findings in a registry.

What's the best way to audition hypotheses before running an experiment?▼

Auditioning hypotheses involves selecting 2–3 candidate hypotheses, choosing a commit boundary, and running a structured evaluation with a pre-registered rubric to ensure rigorous and traceable experiment design.

Related Skills

Explore More Education & Research Skills→
MileniumTickMileniumTick

trade-hypothesis-ideator

Generate ranked falsifiable trade hypotheses with experiment designs and kill criteria.

Community
Advanced
prathamchopra001prathamchopra001

Report Writing

Synthesize research findings into structured JSON discovery reports with themes and evidence.

Community
Intermediate
atholaathola
326

evaluation-framework

Evaluate artifacts using weighted criteria and threshold-based decision rules.

Community
Basic
kofttlcckofttlcc
14

competitive-ads-extractor

Extracts and analyzes competitor ads from Facebook Ad Library and LinkedIn for campaign insights.

Community
Intermediate