camoufox-stealth-browser

Pass

Audited by Gen Agent Trust Hub on Jul 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill requires the installation of the camoufox and curl_cffi Python packages. Furthermore, the scripts/setup.sh script and SKILL.md instructions utilize camoufox.install(), which downloads a custom-patched Firefox browser binary (approximately 700MB) from an external repository during the setup phase.
  • [COMMAND_EXECUTION]: The scripts/setup.sh utility employs sudo dnf install to acquire system dependencies (such as gtk3, libXt, and nss) required for headed browser operation within the isolated distrobox environment.
  • [PROMPT_INJECTION]: The skill facilitates the ingestion of untrusted data from remote web sources, creating a surface for indirect prompt injection.
  • Ingestion points: External content is retrieved via scripts/camoufox-fetch.py and scripts/curl-api.py during web scraping operations.
  • Boundary markers: The skill lacks explicit boundary markers or instructions for the agent to disregard commands embedded within the retrieved data.
  • Capability inventory: The skill scripts have permissions to write the ingested content to the local filesystem and output it to the terminal.
  • Sanitization: There is no evidence of sanitization, filtering, or escaping applied to the fetched HTML or JSON data before it is processed or stored, which could influence downstream agent behavior if the output is interpreted as instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 17, 2026, 04:31 AM