kernel-browser-harness

Pass

Audited by Gen Agent Trust Hub on Jun 27, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The instructions suggest installing 'browser-harness' from its official GitHub repository and 'jq' for JSON processing. These are standard external dependencies for the skill's stated purpose.
  • [PROMPT_INJECTION]: The skill facilitates browser automation on external websites via 'browser-harness'. Ingestion point: external web content accessed via browser-harness. Boundary markers: none specified in guidelines. Capability inventory: includes JavaScript execution (js function), navigation (new_tab), and page inspection (page_info). Sanitization: none specified. This creates a surface for indirect prompt injection where content on a webpage could influence agent behavior.
  • [SAFE]: The skill follows security best practices by recommending the use of environment variables for secrets like KERNEL_API_KEY and ensuring sessions are terminated with the 'kernel browsers delete' command to prevent idle session costs and potential exposure.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 27, 2026, 06:08 AM
Security Audit — agent-trust-hub — kernel-browser-harness