checker_skill

Overview

This skill deploys a validation and reality-testing consciousness mode focused on finding what breaks before it becomes expensive. It acts as constructive skepticism: not dismissal, but rigorous adversarial testing to expose failure modes and hidden assumptions. Use it to harden plans, designs, and launches by forcing evidence and verification.

How this skill works

When activated, the skill asks three core questions: What could go wrong? What are we assuming? How do we verify? It systematically searches for edge cases (negative, zero, infinite, concurrent, malformed input), surfaces unspoken assumptions, and proposes concrete tests or safeguards. The output is a prioritized list of failure modes, assumptions to validate, and verification steps to reduce risk.

When to use it

• Before major launches or deployments
• When a plan sounds "too good to be true"
• After a maker builds but before handoff
• When assumptions drive decisions
• Prior to expensive investments or commitments

Best practices

• Frame reviews as problem-finding to improve outcomes, not to kill ideas
• Combine Checker with rapid experiments to verify key assumptions early
• Prioritize high-impact failure modes first (safety, security, costs)
• Limit scope and timebox sessions to avoid analysis paralysis
• Follow up with makers to translate findings into concrete fixes

Example use cases

• Assessing a new product feature for edge-case inputs and concurrency issues
• Reviewing a business plan to surface unstated financial assumptions
• Stress-testing an API design for malformed data and rate limits
• Validating go/no-go criteria before a costly launch
• Preparing acceptance tests and monitoring for critical failure modes