multi-model-meta-analysis

Extract, deduplicate, and verify cross-model claims against a codebase.

5|2|Updated Jan 20, 2026

Search Tags:#documentation #verification #evidence #code-review #analyze #model-claims

Ready to run

npx skills add https://github.com/petekp/agent-skills --skill multi-model-meta-analysis

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: multi-model-meta-analysis
Source: https://github.com/petekp/agent-skills/tree/main/skills/multi-model-meta-analysis
Command: npx skills add https://github.com/petekp/agent-skills --skill multi-model-meta-analysis

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Synthesize and verify cross-model outputs against a codebase to produce a reliable assessment.

Core Features & Use Cases

Extract discrete claims from multiple model outputs and tag each with its source model.
Deduplicate semantically equivalent claims and create canonical phrasing for clarity.
Verify every factual claim against the actual source code and document evidence.
Resolve conflicts between models and present a reconciled conclusion with supporting evidence.
Generate a structured final report suitable for code reviews and project documentation.

Quick Start

Paste the model outputs and point me to the relevant codebase to generate a verified assessment with evidence.

Frequently Asked Questions about multi-model-meta-analysis

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

FAQPage Schema

How do I verify AI model claims against a codebase?▼

To verify model claims against a codebase, extract discrete assertions from multiple AI outputs, tag them by source, deduplicate semantically equivalent statements, and check each factual claim against the actual source code to document supporting evidence.

What is cross-model output synthesis for code review?▼

Cross-model output synthesis consolidates claims from multiple AI models about code or projects, resolving contradictions and de-duplicating statements to produce a single reconciled conclusion with citations for reliable code review.

How do I reconcile contradictions between different AI models analyzing code?▼

Reconciling contradictions requires extracting claims from each model output, verifying them against the actual source code, resolving conflicts using documented evidence, and generating a structured final report with citations.

What's the best way to consolidate multiple AI code analysis outputs into one report?▼

Consolidating multiple AI code analysis outputs involves extracting discrete claims, deduplicating them into canonical phrasing, verifying facts against the codebase, and generating a structured final report with citations suitable for project documentation.

Do I need to provide source code to verify AI model outputs?▼

Yes, you must point the analysis to the relevant codebase. Verifying factual claims and resolving conflicts requires checking the model outputs against the actual source code to document accurate supporting evidence.