experiment-plan

Generate claim-driven experiment roadmaps with ablation matrices and execution trackers.

1|Updated May 14, 2026
One-click install
npx skills add https://github.com/lix965996-art/MMM --skill experiment-plan-lix965996-art
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: experiment-plan
Source: https://github.com/lix965996-art/MMM/tree/main/resources/app/skills/experiment-plan
Command: npx skills add https://github.com/lix965996-art/MMM --skill experiment-plan-lix965996-art

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps you transform a refined research proposal into a concrete, paper-ready experiment plan that validates specific claims with the right evidence, order, and ablation structure.

Core Features & Use Cases

  • Claim freezing: Converts a method proposal into primary/supporting claims and an anti-claim to rule out confounders.
  • Compact paper storyline: Chooses a minimal set of experiment blocks (main anchor, novelty isolation, simplicity, frontier necessity, and failure analysis) to fit a defensible narrative.
  • Execution-ready roadmap: Produces an experiment block spec plus a realistic run order with milestones, go/no-go gates, and budget/risk estimates.
  • Artifact generation: Writes an experiment plan and an execution tracker into the refine-logs directory.

Quick Start

Use the experiment-plan skill to create a detailed experimental roadmap for your refined method and generate refine-logs/EXPERIMENT_PLAN.md and refine-logs/EXPERIMENT_TRACKER.md.

Frequently Asked Questions about experiment-plan

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

FAQPage Schema
How do I design an experiment plan that validates research claims for a paper?

To design a claim-driven experiment plan, you freeze primary and supporting claims from your method proposal, then map each claim to minimal experiment blocks covering main anchor, novelty isolation, simplicity, frontier necessity, and failure analysis for a defensible narrative.

What is an ablation matrix and how does it isolate novelty in LLM or diffusion research?

An ablation matrix is a structured experiment block that systematically removes or modifies components of your method to isolate which specific design choices drive performance gains. It rules out confounders by testing an anti-claim against your primary research claims.

How do I structure an evaluation protocol with baseline comparisons and run orders?

Structure an evaluation protocol by specifying decisive metrics and setups per experiment block, defining baseline comparisons, and producing a realistic run order with milestones, go/no-go gates, and compute or data budget estimates for each phase.

Can I generate a compute budget and execution tracker for RL-style contributions?

Yes, the experiment plan produces a realistic run order with budget and risk estimates tailored for LLM, VLM, diffusion, and RL-style contributions, outputting an execution tracker that distinguishes must-run from nice-to-have milestones.

What goes into a paper-oriented experiment roadmap for validating research claims?

A paper-oriented experiment roadmap includes frozen primary and supporting claims, an anti-claim to rule out confounders, compact experiment blocks for storyline fit, decisive metrics per block, and a run order with go or no-go gates and compute budgets.

When should I not use a claim-driven experiment planning approach?

This approach is not suitable when you lack a refined research proposal or method idea, as it requires extracting problem anchors and reviewer concerns from existing refine-logs files to produce the experiment plan and execution tracker.