proof-checker-v2

Audit mathematical proofs for missing assumptions and unjustified steps.

2|1|Updated May 7, 2026
One-click install
npx skills add https://github.com/shenmuxing/EtaSkill --skill proof-checker-v2
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: proof-checker-v2
Source: https://github.com/shenmuxing/EtaSkill/tree/main/skill-examples/proof-checker-v2
Command: npx skills add https://github.com/shenmuxing/EtaSkill --skill proof-checker-v2

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires deepseek-agent, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill performs an independent adversarial review of existing proofs, helping catch hidden assumptions, missing steps, quantifier mistakes, and unsupported claims before a proof is accepted.

Core Features & Use Cases

  • Proof Auditing: Reviews theorem, lemma, proposition, and proof artifacts in Markdown, LaTeX, or proof logs.
  • Adversarial Verification: Checks for illegal interchanges, dependency gaps, circularity, scope overclaims, and other structural errors.
  • Workflow Support: Fits as a reviewer step in larger proof workflows where a separate verification pass is needed.
  • Use Case: Ask it to check a completed proof draft and return a verdict with issues, counterexample attempts, and minimal repairs.

Quick Start

Use the proof-checker-v2 skill to audit this proof and return a structured verdict with the most serious gap, if any.

Frequently Asked Questions about proof-checker-v2

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

FAQPage Schema
How do I audit a mathematical proof for hidden assumptions and missing steps?

You can audit a mathematical proof by running an adversarial review pass that locates missing assumptions, unjustified steps, and hidden dependency errors. This process applies rubric-based checking to return a structured verdict with evidence-backed issue reporting.

What is adversarial proof verification and how does it catch structural errors?

Adversarial proof verification is an independent review mechanism that checks for illegal interchanges, dependency gaps, circularity, and scope overclaims. It applies rubric-based checking to identify structural errors that might invalidate a completed theorem or lemma proof.

Can I verify LaTeX or Markdown proof artifacts for circularity and scope overclaims?

Yes, you can verify LaTeX and Markdown proof artifacts for circularity and scope overclaims. The adversarial review process supports theorem, lemma, proposition, and proof logs in these formats to locate structural errors and unsupported claims.

Does proof auditing work with DeepSeek for external theorem verification?

Yes, proof auditing supports external DeepSeek review routing for theorem verification. This integration enables structured proof review with rubric-based checking, routing completed proofs to DeepSeek for an independent correctness pass and evidence-backed issue reporting.

What is the best way to check a completed proof draft for quantifier mistakes?

The best way to check a completed proof draft for quantifier mistakes is to run an independent adversarial review pass. This approach catches hidden assumptions, quantifier errors, and unsupported claims by applying rubric-based verification before the proof is accepted.

When should I not use automated proof checking for mathematical theorems?

Automated proof checking may not be suitable when a proof artifact lacks structured formatting in Markdown, LaTeX, or proof logs, or when the verification workflow cannot support external DeepSeek review routing and rubric-based checking for evidence-backed issue reporting.