postiz

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md and Integration Tools Workflow explicitly instruct the agent to call integrations:trigger / integrations:settings to fetch dynamic data from third‑party platforms (e.g., Reddit flairs, YouTube playlists, LinkedIn companies) and to read and use those outputs when constructing posts (see "Pattern 1", "Integration Tools Workflow" and examples), so untrusted/user-generated content from public sites can directly influence tool use and next actions.