The Agent Skills Directory

[SAFE]: The skill implements a robust security validation framework. It uses local libexec scripts for sandboxed execution of proof-of-concept code and incorporates Z3 SMT solver integration for formal verification of path conditions. The instructions emphasize human review, strict schema validation, and harmless PoC development. All external references, such as the exploitation-validator methodology on GitHub, point to the official repository of the skill's author, gadievron. The framework follows least-privilege principles by restricting network access and filesystem writes during the validation process.

exploitability-validation