visa-doc-translate
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes text extracted via OCR from user-provided images of sensitive documents such as passports or bank statements. This creates a surface where malicious instructions embedded within these documents could potentially influence the agent's behavior. Ingestion points: User-supplied image file paths in SKILL.md. Boundary markers: Absent; there are no instructions to ignore or delimit embedded text extracted from images. Capability inventory: Shell command execution (sips), package installation (pip, brew), and Python script generation/execution in SKILL.md. Sanitization: Absent; the skill translates and reformats extracted text without validation or filtering.
- [DYNAMIC_EXECUTION]: The skill instructions direct the agent to dynamically generate and execute a Python script using the PIL and reportlab libraries to create the final PDF document.
- [COMMAND_EXECUTION]: The skill utilizes the macOS sips utility to perform image conversion, which involves executing shell commands with user-supplied file paths.
Audit Metadata