dwarf-expert

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill provides authoritative guidance on DWARF debug information analysis. It directs users to official resources like dwarfstd.org and well-known, established open-source repositories for libdwarf, pyelftools, and gimli. The instructions are consistent with legitimate security research and development workflows, and no patterns of obfuscation, exfiltration, or unauthorized command execution were detected.\n- [INDIRECT_PROMPT_INJECTION]: The skill involves the analysis of external, potentially untrusted binary files. This creates a surface for indirect prompt injection if an attacker crafts malicious DWARF metadata intended to influence the agent's interpretation of tool outputs.\n
  • Ingestion points: Compiled binaries analyzed using llvm-dwarfdump, readelf, or scripts utilizing pyelftools (SKILL.md).\n
  • Boundary markers: Not utilized; the agent processes output directly from the command-line tools.\n
  • Capability inventory: The skill utilizes Bash for executing diagnostic tools and WebFetch for standard documentation lookups.\n
  • Sanitization: Not explicitly discussed; the skill relies on the structured output and parsing logic of standard binary utilities.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 01:16 PM
Security Audit — agent-trust-hub — dwarf-expert