browse-ai

Pass

Audited by Gen Agent Trust Hub on Mar 28, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill facilitates the extraction of structured data from external websites, which introduces a surface for indirect prompt injection where instructions embedded in scraped content could influence agent behavior.
  • Ingestion points: Website data retrieved via RunExtraction, RunTask, and GetTask actions in SKILL.md.
  • Boundary markers: The instructions do not specify the use of delimiters or 'ignore embedded instructions' warnings when handling scraped data.
  • Capability inventory: The skill allows the agent to execute shell commands via the Membrane CLI (membrane action run, membrane request) and perform network operations.
  • Sanitization: No explicit sanitization or content validation of the scraped data is mentioned in the skill body.
  • [EXTERNAL_DOWNLOADS]: The skill identifies the installation of a command-line utility from the vendor as a prerequisite for managing the integration.
  • Evidence: Recommends global installation of the @membranehq/cli package via npm in SKILL.md.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 28, 2026, 07:41 PM