azure-upgrade

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly instructs the agent to fetch and read external public documentation (e.g., "For each migration guide you recorded during migration: Fetch and read the full content of the guide URL" in references/languages/java/INSTRUCTION.md and numerous aka.ms / raw.githubusercontent.com / GitHub links), meaning the agent will ingest open/public third‑party content that can influence upgrade decisions and actions.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly requires fetching external guide/POM content at runtime (e.g., "fetch https://raw.githubusercontent.com/Azure/azure-sdk-for-java/main/sdk/boms/azure-sdk-bom/pom.xml" and other raw GitHub/guide URLs) and uses that content to drive migration validation and instructions, so this is a runtime external dependency that can directly control the agent's behavior.