visa-doc-translate
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes shell commands for image processing (
sips) and environment setup (pip install) automatically without seeking user confirmation. - [DYNAMIC_EXECUTION]: The skill generates and executes a Python script at runtime to produce the final PDF document. While intended for formatting, this involves assembling code from processed data.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data extracted from images via OCR. This content is fed directly into the agent's context for translation without sanitization or boundary markers.
- Ingestion points: User-supplied image files (PNG, HEIC, JPG, etc.) containing visa document information.
- Boundary markers: None identified; the instructions do not define delimiters for OCR-extracted text or warn the agent against following instructions found within the images.
- Capability inventory: Shell command execution (
sips), automated package installation, and execution of generated Python scripts. - Sanitization: None identified; OCR output is processed directly as text to be translated and formatted.
Audit Metadata