The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to fetch documentation from learn.microsoft.com using tools like fetch_webpage. Processing external, untrusted content (even from trusted domains) and incorporating it into the agent's context creates a surface for indirect prompt injection where instructions embedded in the web content could influence the agent's behavior.
Ingestion points: Content retrieved via mcp_microsoftdocs:microsoft_docs_fetch and fetch_webpage as described in SKILL.md.
Boundary markers: None explicitly defined to delimit fetched content or warn the agent about potential instructions within the data.
Capability inventory: The agent has capabilities for network access and local file reading (read_file).
Sanitization: No explicit sanitization or validation of the fetched Markdown content is specified in the skill instructions.
[EXTERNAL_DOWNLOADS]: The skill fetches documentation from learn.microsoft.com and directs users to the official MicrosoftDocs GitHub repository for tool installation. These operations target well-known and established service providers.

azure-personalizer