The Agent Skills Directory

[SAFE]: The skill provides conceptual and code-level guidance for static analysis tasks such as pointer analysis, taint tracking, and dependency graphing using the LLVM framework.
[EXTERNAL_DOWNLOADS]: The skill references a external repository (gmh5225/awesome-llvm-security) to fetch curated lists of security tools and research. This is a well-known, community-maintained collection of information relevant to the skill's primary purpose. No execution of remote code is performed.
[COMMAND_EXECUTION]: Code snippets provided are for C++ development with the LLVM library and are intended for implementation within a static analysis tool, not for direct execution by the agent.
[INDIRECT_PROMPT_INJECTION]: While the skill discusses processing external code (static analysis), it does not define a runtime data ingestion surface that would allow for indirect prompt injection against the agent itself.

static-analysis