The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill instructions direct the agent to 'Write a script or test that runs the real code' to prove a change is safe. This is a core part of the skill's intended purpose for code auditing and is confined to the codebase being analyzed.- [SAFE]: The skill includes explicit instructions to 'strip anything private before it goes anywhere public,' demonstrating a privacy-conscious design. No remote code execution from untrusted sources, data exfiltration, or obfuscation patterns were detected.

blast-radius