math-cryptography

Pass

Audited by Gen Agent Trust Hub on Sep 7, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill is restricted to generating text-based reasoning artifacts. It does not request or use any tools that interact with the host system or external networks.
  • [INDIRECT_PROMPT_INJECTION]: The skill accepts user-provided text as input for analysis.
  • Ingestion points: Cryptographic problem description and inputs in the skill body.
  • Boundary markers: Not present.
  • Capability inventory: The skill does not use any executable tools, perform network access, or write to the file system beyond creating the markdown artifact.
  • Sanitization: No specific sanitization measures are implemented for the input text.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 7, 2026, 02:59 AM
Security Audit — agent-trust-hub — math-cryptography