facebook-marketplace-message-sender

Pass

Audited by Gen Agent Trust Hub on Aug 15, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [SAFE]: The skill mandates a user-provided approval_token for each outbound message, ensuring that no message can be sent without explicit human verification and authorization.
  • [COMMAND_EXECUTION]: The skill utilizes the browser-harness tool to automate browser tasks, focusing on navigating to Facebook Marketplace and interacting with its messaging interface.
  • [DYNAMIC_EXECUTION]: Embedded Python scripts are executed through shell heredocs into the browser-harness to perform complex UI interactions like clearing text and verifying message bubbles.
  • [DATA_EXPOSURE]: Thread state is managed locally in the reports/threads/ directory, used to track outbound message history and conversation status.
  • [SAFE]: The skill incorporates a content validation step via facebook-marketplace-safety-guard, which checks the proposed message against safety policies before it is sent.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 15, 2026, 01:26 PM
Security Audit — agent-trust-hub — facebook-marketplace-message-sender