heuristics
Pass
Audited by Gen Agent Trust Hub on Jun 15, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted data from user-provided files, including documents, code, and images. Maliciously crafted instructions within these source files could attempt to influence the agent's behavior during the classification or storage process.
- Ingestion points: User-provided file paths and folder contents as described in the 'Extraction Routing' and 'Mixed Folders' sections.
- Boundary markers: The instructions do not specify the use of delimiters or 'ignore embedded instructions' warnings when the agent processes the extracted content.
- Capability inventory: The skill utilizes file system reading, optional vision tools for image processing, the 'markitdown' CLI for document conversion, and the '@historian' memory API for persistent storage.
- Sanitization: The skill includes a 'What Not To Store' section that explicitly directs the agent to identify and exclude secrets, credentials, and private personal data, which is a significant mitigating factor against data exposure.
Audit Metadata