symbolic-execution-tools

Pass

Audited by Gen Agent Trust Hub on Jul 21, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill recommends installing several well-known security analysis libraries via pip, including angr, z3-solver, unicorn, capstone, and keystone-engine. These are standard tools for the skill's stated purpose of binary reversing and symbolic execution.
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it is designed to analyze untrusted binary files provided by the user (e.g., CTF challenges).
  • Ingestion points: The skill uses angr.Project('./challenge') in both SKILL.md and ANGR_COOKBOOK.md to load external binaries into the analysis environment.
  • Boundary markers: Absent. There are no specific instructions or delimiters guiding the agent to ignore potentially malicious instructions embedded within the analyzed binary data.
  • Capability inventory: The skill utilizes frameworks that perform full CPU emulation, memory manipulation, and symbolic execution.
  • Sanitization: No sanitization is performed on the binary content before it is processed by the emulation and solver engines.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 21, 2026, 01:24 PM
Security Audit — agent-trust-hub — symbolic-execution-tools