visa-doc-translate
Pass
Audited by Gen Agent Trust Hub on Jul 17, 2026
Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [DATA_EXFILTRATION]: The skill is designed to process highly sensitive data, including passports, ID cards, bank deposit certificates, and employment records. While no active exfiltration to external domains was detected, the processing of such high-value personally identifiable information (PII) increases the security risk profile if the agent's environment is compromised.
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface. It ingests untrusted text data via OCR from user-provided document images (SKILL.md, Step 3) and incorporates this content into the agent's context for translation. There are no boundary markers or instructions to ignore embedded commands within the extracted text. Since the agent has the capability to execute shell commands and scripts, a malicious document containing hidden instructions could potentially trigger unauthorized actions.
- [COMMAND_EXECUTION]: The skill uses dynamic execution by generating and running Python scripts locally to build PDF documents (SKILL.md, Step 5). It also utilizes shell commands like
sipsfor image conversion. The interpolation of user-provided filenames directly into these commands and scripts without explicit sanitization could lead to local command injection if a file is maliciously named.
Audit Metadata