patsnap-chemical-molecular
Pass
Audited by Gen Agent Trust Hub on Apr 23, 2026
Risk Level: SAFENO_CODEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill consists entirely of documentation and configuration examples. No executable scripts, binaries, or active code are included.
- [EXTERNAL_DOWNLOADS]: The skill instructs the agent to connect to an external MCP server at connect.patsnap.com. This is the official infrastructure for the PatSnap service and represents normal functionality.
- [PROMPT_INJECTION]: The skill describes an interaction model where the agent ingests data from external chemical databases and patent records, creating a surface for potential indirect prompt injection. 1. Ingestion points: Data returned from the connect.patsnap.com MCP server (patent text, chemical metadata). 2. Boundary markers: No specific delimiters or safety instructions are defined in the SKILL.md to distinguish data from instructions. 3. Capability inventory: The skill does not request any dangerous capabilities such as file system writes, network exfiltration, or shell execution. 4. Sanitization: No sanitization logic is provided in the documentation to filter external content.
Audit Metadata