steelman

Identify user claims, gather repository evidence, and issue a formal verdict.

3|1|Updated Mar 23, 2026

Search Tags:#evidence #code-review #design-review #risk-analysis #steelman #counter-argument

Ready to run

npx skills add https://github.com/malinskibeniamin/skills --skill steelman

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: steelman
Source: https://github.com/malinskibeniamin/skills/tree/main/steelman
Command: npx skills add https://github.com/malinskibeniamin/skills --skill steelman

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill helps you counter user assertions with evidence to prevent unexamined agreement, ensuring high-stakes decisions are grounded in analysis rather than confirmation bias.

Core Features & Use Cases

Identify claim: Restate the user assertion to establish scope.
Evidence gathering: Collect repository signals, tests, docs, and references to challenge the claim.
Steelman opposite: Present the strongest, well-supported counter-argument with precise references.
Verdict & guidance: Deliver a clear conclusion (Confirmed, Contradicted, or Mixed) with recommended next steps.

Quick Start

When a user makes a claim, restate the claim, gather repo evidence, and present the strongest opposing argument along with a verdict.

Frequently Asked Questions about steelman

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

FAQPage Schema

What is steelmanning in code review and design evaluation?▼

Steelmanning in code review is the practice of identifying a developer's claim and constructing the strongest possible counter-argument using repository evidence to prevent unexamined agreement and confirmation bias.

How do I challenge an architecture assertion using repository evidence?▼

To challenge an architecture assertion, restate the claim to establish scope, gather repository signals and tests as evidence, present the strongest opposing argument with references, and deliver a formal verdict of Confirmed, Contradicted, or Mixed.

Can I use this approach to test high-stakes debugging claims?▼

Yes, you can use this approach to test high-stakes debugging claims by applying structured claim identification and evidence gathering to challenge premises where wrong conclusions incur significant costs.

What is the best way to evaluate risk analysis assumptions in software engineering?▼

The best way to evaluate risk analysis assumptions is to construct a well-supported counter-argument with precise references, ensuring high-stakes decisions are grounded in documented analysis rather than unexamined agreement.

When should I avoid presenting a counter-argument during a design review?▼

You should avoid presenting a counter-argument during routine or low-stakes design reviews where the cost of wrong premises is minimal, as the formal evidence gathering and verdict process is intended for high-stakes assertions.