target-discovery
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external, untrusted data which serves as a potential surface for indirect prompt injection techniques.\n
- Ingestion points: The skill accepts chemical SMILES strings from users and retrieves external patent and literature data through specialized search tools (SKILL.md).\n
- Boundary markers: The instructions do not specify the use of delimiters or ignore-instructions to isolate retrieved scientific data from the agent's core instructions.\n
- Capability inventory: The skill coordinates a wide range of analytical tools, including structure searches, toxicity predictions, and report generation (PPT/PDF).\n
- Sanitization: Basic SMILES format verification is recommended before processing, though specific sanitization for text retrieved from scientific databases is not detailed.
Audit Metadata