The Agent Skills Directory

[SAFE]: The skill functions as a structured index for Azure Repos documentation, mapping technical topics to official Microsoft Learn URLs. It does not contain executable code.
[EXTERNAL_DOWNLOADS]: The skill defines patterns for fetching documentation from learn.microsoft.com using tools like mcp_microsoftdocs:microsoft_docs_fetch. These requests target authoritative sources associated with the author ('microsoftdocs') and include benign tracking parameters (e.g., from=learn-agent-skill).
[PROMPT_INJECTION]: The instructions include 'IMPORTANT for Agent' directives. These are used to guide the model on how to navigate the skill's content and tool usage (e.g., checking version freshness or using specific fetching tools). They do not attempt to bypass safety guardrails or override core agent behavior.
[COMMAND_EXECUTION]: While the skill mentions various CLI commands (e.g., tf.exe, git) and integration patterns, these are provided as documentation and examples for the user, not for automatic execution by the agent in a high-risk manner.

azure-repos