privacy-mask

Pass

Audited by Gen Agent Trust Hub on Jul 3, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill instructs the agent to use a local CLI tool (privacy-mask) to redact sensitive information (PII) from images. All processing is explicitly stated as local and offline.
  • [EXTERNAL_DOWNLOADS]: The skill mentions the installation of the privacy-mask package via pip and system dependencies like tesseract via brew or apt. These are standard installation procedures for a legitimate tool and do not involve untrusted or suspicious sources.
  • [COMMAND_EXECUTION]: The skill uses standard command-line execution for the privacy-mask tool. It provides parameters for masking and dry-runs, which are consistent with the tool's documented purpose.
  • [DATA_EXFILTRATION]: There is no evidence of network operations or data exfiltration. The instructions explicitly forbid sending unmasked images to external APIs or cloud services, emphasizing local processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 3, 2026, 01:06 PM
Security Audit — agent-trust-hub — privacy-mask