mom-test

What it does

The mom-test skill is a utility designed to evaluate AI agent responses against the rigorous standards of a critical parent. It provides immediate feedback on tone, logic, and helpfulness, simulating the scrutiny often applied to ensure quality and safety in outputs.

When to use it

• Before deploying an agent to sensitive environments where tone and reliability are paramount.
• To identify hallucinations or logical gaps in complex reasoning tasks.
• During iterative development to refine conversational nuances and adherence to constraints.
• When you need a quick, informal audit of an agent's performance without formal testing frameworks.

Key capabilities

• Simulates critical feedback loops typical of parental oversight.
• Assesses response quality based on tone, clarity, and accuracy.
• Highlights areas where the agent may be too verbose, dismissive, or incorrect.

Example prompts

• "Run a mom-test on this draft email to my boss to check if it sounds professional and confident."
• "Act as a strict parent and critique this AI's explanation of quantum physics for clarity and accuracy."
• "Perform a mom-test on the following code snippet to ensure it is bug-free and well-commented before I share it."

Tips & gotchas

This skill relies on qualitative judgment rather than automated metrics, so results are subjective and depend on the specific context provided. It is best used as a supplementary check alongside formal testing tools for critical applications.