ucp-schema-authoring

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

• Third-party content exposure detected (high risk: 1.00). The SKILL.md explicitly instructs the agent to fetch live spec pages from ucp.dev (e.g., "Fetch live spec" and https://ucp.dev/specification/reference/) and to fetch base/extension schemas from schema URIs, meaning it will read public third-party content that directly influences schema composition and subsequent actions.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill explicitly instructs fetching live spec and schema files at runtime (e.g., https://ucp.dev/specification/reference/ and https://ucp.dev/schemas/shopping/checkout.json), and those fetched JSON schemas are required to drive schema composition/validation and thus directly control the agent's behavior when authoring schemas.