microsoft-skill-creator

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

• Third-party content exposure detected (high risk: 0.85). The required workflow uses Learn MCP tools like microsoft_docs_fetch(url=...) to retrieve full page content from Microsoft Learn at runtime, which is outsider-authored web text that can be ingested into the agent’s LLM context.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly calls microsoft_docs_fetch at runtime to load external documentation pages (e.g., "https://learn.microsoft.com/semantic-kernel/overview/" and "https://learn.microsoft.com/en-us/semantic-kernel/frameworks/agent/agent-memory"), so remote content is fetched during execution and can directly control the agent's prompts/instructions.