The Agent Skills Directory

Indirect Prompt Injection Surface: The skill is designed to process 'candidate findings' generated by discovery tools or provided by users. Because these findings originate from untrusted external sources, they represent a potential attack surface where malicious data could attempt to influence the agent's validation workflow. The skill includes protective measures such as requiring a structured validation rubric, enforcing non-interactive command execution, and mandating explicit evidence for every finding assessment.
Dynamic Execution for Security Validation: To perform its intended function, the skill utilizes dynamic execution methods, including compiling debug variants of code, generating Proof-of-Concept (PoC) scripts, and invoking debuggers like gdb or lldb and memory analyzers such as valgrind or ASan. These capabilities are necessary for providing high-confidence security assessments and are limited by instructions to use short, bounded, and non-interactive command sequences to maintain safety.

validation