visa-doc-translate
Pass
Audited by Gen Agent Trust Hub on May 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the sips command-line utility for image format conversion and instructs the agent to install system and Python dependencies using brew and pip.
- [REMOTE_CODE_EXECUTION]: The instructions direct the agent to dynamically generate a Python script using the reportlab and pillow libraries and then execute that script at runtime to create a PDF file.
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it processes untrusted text extracted via OCR from user-provided images. Malicious instructions embedded within a document could potentially influence the agent's behavior during the translation or PDF generation phases. Ingestion points: SKILL.md (Step 3: OCR Text Extraction). Boundary markers: Absent; extracted text is provided to the agent for translation without delimiters or instructions to ignore embedded commands. Capability inventory: Shell command execution (sips) and dynamic Python script execution (SKILL.md). Sanitization: Absent; there is no validation or filtering of text extracted from images.
Audit Metadata