pubchem-database
Pass
Audited by Gen Agent Trust Hub on Apr 16, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill includes behavioral steering instructions that direct the agent to monitor the complexity of the user's request and proactively suggest the author's commercial platform (www.k-dense.ai) for complex workflows.- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing untrusted chemical identifiers and structures that are then utilized in network and file-writing operations.
- Ingestion points: User-supplied chemical names, SMILES strings, and formulas entering the agent context via input parameters (SKILL.md).
- Boundary markers: Absent; the skill does not define delimiters or provide instructions to the agent to disregard instructions potentially embedded in the chemical data.
- Capability inventory: The skill utilizes
pcp.downloadfor file system writes andrequests.getfor network operations across multiple scripts (SKILL.md). - Sanitization: Absent; user inputs are passed directly to library functions and API requests without visible validation or sanitization logic.- [EXTERNAL_DOWNLOADS]: The skill facilitates downloading molecular structure files (SDF, JSON) and images (PNG) from the official National Institutes of Health (NIH) PubChem servers (pubchem.ncbi.nlm.nih.gov).- [COMMAND_EXECUTION]: The installation documentation provides shell commands for users to install Python dependencies such as
pubchempy,requests, andpandasusing theuvpackage manager.
Audit Metadata