code-pipelines

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches and ingests untrusted third-party content (web research results from Tavily and Exa in reference/enrichment-apis.md and pipelines/outbound/cold-outreach.ts, plus inbound emails read from Gmail in pipelines/conversational/email-reply-handler.ts), and that content is directly used by the AI to generate replies, outreach, scheduling, and other actions—allowing arbitrary external text to influence tool use and next actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill requires and calls the Personize API at https://app.personize.ai (e.g., personize.agents.run → POST /api/v1/agents/:id/run), which at runtime fetches and executes server-stored agent instructions and prompts that directly control agent behavior.