nice-try-gpt
Analyze an authorized CTF challenge, reproduce its intended solve, identify cheap LLM shortcuts, apply minimal human-friendly transformations, and verify the result end-to-end. Use when reviewing or adapting CTFs to reduce pattern-matching shortcuts without materially increasing human difficulty.
UncategorizedView skill →
