rsa-attack-techniques
Pass
Audited by Gen Agent Trust Hub on Nov 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides example commands for running established security tools such as RsaCtfTool, yafu, and msieve. These tools are used locally for legitimate security research and CTF purposes.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted cryptographic parameters (n, e, c). This constitutes a potential attack surface, although the risk is minimal given the mathematical nature of the inputs and the skill's intended use.\n
- Ingestion points: RSA parameters (n, e, c) and public key files (.pem).\n
- Boundary markers: None identified; inputs are handled as integers or strings within mathematical scripts.\n
- Capability inventory: Implementation of cryptanalysis algorithms and local execution of factorization tools.\n
- Sanitization: Standard mathematical parsing for cryptographic variables.
Audit Metadata