posthog-inbound-leads

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

• Third-party content exposure detected (high risk: 0.80). The skill's required workflow explicitly mandates fetching and interpreting every URL included in the draft email using web_fetch (SKILL.md — "Step 4: Validate All URLs") and also ingests whatever lead content the TAE provides (Core Workflow Step 1), meaning the agent will retrieve and act on arbitrary, untrusted third‑party pages and user-provided content as part of its decision-making.