action-builder-skill

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

• Third-party content exposure detected (high risk: 0.80). The skill instructs the agent to fetch and read external API documentation and user-provided API reference URLs (see "Using User-Provided Values" → "API reference URL → Fetch for schema details" and "When API Docs Don't Render" mentioning WebFetch), which means it would ingest arbitrary public third-party web content as part of its workflow.