dynamic-instrumentation
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches a curated list of LLVM security tools and projects from the author's repository on GitHub (gmh5225/awesome-llvm-security).
- [INDIRECT_PROMPT_INJECTION]:
- Ingestion points: The skill instructs the agent to fetch external resource data from a remote GitHub README file.
- Boundary markers: None identified in the instruction to fetch external content.
- Capability inventory: The skill contains implementation patterns for memory access tracking, system call monitoring, and runtime binary instrumentation.
- Sanitization: No explicit sanitization or validation is performed on the content retrieved from the external URL.
Audit Metadata