skills.rest
TabithaFanny
TabithaFannyCOMMUNITY
Explore all skills by this author →
GitHub

experiment-audit

Verify experimental result integrity by detecting fraud patterns with cross-model review.

Updated May 29, 2026
Search Tags:#experiment analysis#fraud detection#result verification#cross-model review#experiment integrity
One-click install
npx skills add https://github.com/TabithaFanny/ThesisX --skill experiment-audit-tabithafanny
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: experiment-audit
Source: https://github.com/TabithaFanny/ThesisX/tree/main/skills_imported/aris/skills/experiment-audit
Command: npx skills add https://github.com/TabithaFanny/ThesisX --skill experiment-audit-tabithafanny

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires mcp__codex__codex, mcp__codex__codex-reply, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill provides a robust integrity check for experiments, using cross-model review to detect common fraud patterns and ensure the accuracy of reported results.

Core Features & Use Cases

  • Fraud Detection: Checks for fake ground truth, score normalization fraud, phantom results, and insufficient scope.
  • Cross-Model Review: Uses GPT-5.4 to analyze code and results without interference from the executor.
  • Use Case: When conducting experiments, especially after completion, use this Skill to ensure the integrity of the results before making any claims.

Quick Start

Run the experiment-audit skill with the path to your experiment directory.

Frequently Asked Questions about experiment-audit

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

FAQPage Schema
How do I verify experiment integrity and detect result fraud?▼

To verify experiment integrity, cross-model review checks evaluation code and result files to detect fake ground truth, score normalization fraud, and phantom results before you report findings.

What is cross-model review for experiment analysis?▼

Cross-model review is an evaluation method that uses a separate model like GPT-5.4 to analyze experiment code and results independently, preventing executor interference and ensuring reported accuracy.

What files do I need to prepare for an experiment audit?▼

You need to provide access to your experiment code, evaluation scripts, result files, and configuration files so the review process can accurately identify fraud patterns and scope issues.

How does cross-model review detect score normalization fraud?▼

Cross-model review detects score normalization fraud by analyzing evaluation scripts and configuration files to identify manipulated scoring metrics and phantom results within your experiment directory.

When should I run an experiment audit on my results?▼

You should run an experiment audit after completing your experiment but before making any claims, ensuring the integrity of your results by checking for fake ground truth and insufficient scope.

Can I use experiment audit for result verification on large scale experiments?▼

Result verification through cross-model review analyzes provided code, scripts, and files to identify insufficient scope, ensuring your experiment claims remain reliable regardless of the project scale.

Related Skills

Explore More Software Engineering Skills→
zhuyingqinzhuyingqin
1

experiment-bridge

Implement training and evaluation code from EXPERIMENT_PLAN.md and deploy to GPUs.

Community
Advanced
goupup-aigoupup-ai
2

experiment-audit

Audit machine learning experiment results for fraudulent reporting patterns.

Community
Advanced
zeronx798zeronx798

requesting-code-review

Scan staged git changes for security issues and quality regressions.

Community
Advanced
dogekikidogekiki

experiment-audit

Audit experimental integrity by cross-referencing scripts, results, and claims against fraud patterns.

Community
Advanced