verify

Verify in-repo payloads via cross-model reviewers and update prosecution documents.

1|Updated Apr 28, 2026
One-click install
npx skills add https://github.com/recallnet/polymarket-cross-sectional-momentum --skill verify-recallnet
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: verify
Source: https://github.com/recallnet/polymarket-cross-sectional-momentum/tree/main/.agents/skills/verify
Command: npx skills add https://github.com/recallnet/polymarket-cross-sectional-momentum --skill verify-recallnet

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Gate-level cross-model verification to ensure gate flows are validated by a reviewer from a different model family and to close the fix-and-re-review loop before releasing verdict.

Core Features & Use Cases

  • Cross-model reviewer pairing to enforce diversity and reduce blind spots.
  • Payload-driven verification that writes findings into the prosecution document.
  • Fix-and-re-review loop with traceable verdicts and updates across gates.

Quick Start

Invoke verify with a payload and gate to start the cross-model review workflow.

Frequently Asked Questions about verify

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

FAQPage Schema
What is cross-model gate verification and why use different model families?

Cross-model gate verification validates gate flows by using a reviewer from a different model family to reduce blind spots and enforce diversity. It ensures payload-driven checks are independently reviewed before releasing a final verdict.

How do I run a fix-and-re-view loop for prosecution document updates?

To run a fix-and-re-view loop, execute the oracle with a composed prompt, capture findings, and update the prosecution document. Repeat this process until no LANDED findings remain to achieve a traceable verdict.

How do I compose an oracle prompt from a payload's SKILL.md for review?

You compose the oracle prompt by extracting the protocol from the payload's SKILL.md. Select reviewers from different model families, then execute the oracle to apply cross-model verification against the in-repo payload.

Does cross-model verification work without external dependencies for gate reviews?

Yes, cross-model verification operates without external dependencies. It applies the protocol defined internally to select reviewers, compose the oracle prompt, and generate prosecution-doc updates until no LANDED findings remain.

What are the limitations of using a single model family for gate review?

Using a single model family for gate review introduces blind spots and lacks diversity. Cross-model verification addresses this by enforcing reviewer pairing from different model families to close the fix-and-re-view loop before releasing a verdict.