dogfood

Pass

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill functions by ingesting and processing content from external, untrusted websites during QA sessions. This creates a surface where embedded instructions in target applications could potentially influence agent behavior.\n
  • Ingestion points: Data entering the context via camofox navigate, camofox snapshot, and camofox console commands (SKILL.md).\n
  • Boundary markers: The instructions do not define specific delimiters or directives to ignore instructions within the target application's content.\n
  • Capability inventory: The skill utilizes camofox for browser interaction (click, type, navigate) and file system operations for saving trace files (SKILL.md).\n
  • Sanitization: There are no explicit instructions for sanitizing or escaping content retrieved from the target application before it is processed by the agent.\n- [SAFE]: The skill uses camofox-browser, a legitimate tool for anti-detection browser automation. It captures session data and saves traces to a local directory (~/.camofox/traces/) for later analysis. Visual evidence is captured using standard annotation and snapshot tools.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 16, 2026, 08:11 PM
Security Audit — agent-trust-hub — dogfood