permutive

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (low risk: 0.10). The skill reads public pages and search results from permutive.com via an MCP server, which constitutes ingesting content from an external website (low risk).

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (medium risk: 0.30). The skill connects to a remote third-party MCP server hosted at unbrowse.ai to dynamically fetch tools and instructions at runtime, representing an unofficial operational dependency with unknown/third-party publisher provenance.