Stats
Actions
Tags
From ice-framework
Review AI output critically before accepting it. Use when the user invokes /verify or wants to check AI-generated content for errors, assumptions, and claims requiring validation.
How this skill is triggered — by the user, by Claude, or both
Slash command
/ice-framework:verifyThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
After generating output, produce a structured checklist that gives the **human** a clear list of things to check. Your role is not to verify your own output — you cannot do that reliably — but to make verification by the human faster and more systematic.
After generating output, produce a structured checklist that gives the human a clear list of things to check. Your role is not to verify your own output — you cannot do that reliably — but to make verification by the human faster and more systematic.
Fluent text is not accurate text. The more polished and confident AI output appears, the less critically humans tend to review it. This skill exists to break that pattern by surfacing what needs human attention, not by replacing human judgement.
When this skill is invoked, respond with:
Claims requiring human verification:
Potential technical issues:
Assumptions I made:
What a domain expert should review:
Confidence assessment:
This checklist is a prompt for the human, not a verification of the AI output. Generating a checklist is not the same as checking anything. The human must do the actual verification. If the user proceeds without checking, that is their choice — but they should make it consciously, not by default.
Guides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.
npx claudepluginhub topsy31/ice --plugin ice-framework