agents-sdk

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

• Third-party content exposure detected (high risk: 0.90). The skill's SKILL.md explicitly instructs the agent to "Fetch current docs from https://github.com/cloudflare/agents/tree/main/docs" (public GitHub) and the references (e.g., Code Mode and MCP integration) also show fetching/ingesting external MCP servers, emails, and outbound URLs, so the agent will read untrusted, user-generated or public third-party content that can influence its actions.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly instructs fetching runtime documentation from https://github.com/cloudflare/agents/tree/main/docs before implementing, meaning this external URL is fetched at runtime and its content would directly guide/alter agent implementation and behavior.