pathology-koan-generator
Generate diagnostic koans to test reasoning boundaries and edge cases.
When & Why to Use This Skill
The Pathology Koan Generator is a specialized diagnostic tool designed to probe the reasoning limits and edge cases of AI models. By generating complex, paradoxical scenarios known as 'koans,' this skill helps developers and AI researchers identify logical vulnerabilities, cognitive biases, and boundary conditions in Large Language Models (LLMs) and autonomous agents, ensuring more robust and reliable AI performance.
Use Cases
- Model Stress Testing: Generate intricate logical puzzles to determine the exact point where an AI's reasoning chain fails or begins to hallucinate.
- Comparative Benchmarking: Use standardized diagnostic koans to evaluate and compare the reasoning depth and problem-solving capabilities of different model versions or prompt engineering strategies.
- Safety & Robustness Auditing: Proactively discover edge cases in agent behavior to ensure robust performance and adherence to logic in unpredictable or adversarial real-world environments.
| name | pathology-koan-generator |
|---|---|
| description | Generate diagnostic koans to test reasoning boundaries and edge cases. |
Instructions
- Initialize pathology-koan-generator operational context
- Execute primary protocol actions
- Validate results and generate output
Examples
- "Execute pathology-koan-generator protocol"
- "Run pathology koan generator analysis"