state-directory-manager

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

• Third-party content exposure detected (high risk: 0.80). The skill includes examples that run arbitrary commands and fetch open-web content (e.g., get_with_cache with an eval of "curl -s https://api.example.com/data" and the "Usage in Scripts" curl example), so the agent can ingest and read untrusted third-party responses as part of its workflow.