The Agent Skills Directory

[SAFE]: No malicious code patterns, obfuscation, or persistence mechanisms were detected in the skill files or metadata.
[EXTERNAL_DOWNLOADS]: The skill fetches technical reference content from learn.microsoft.com. This is a well-known and established service, and the use case is limited to information retrieval for user support.
[PROMPT_INJECTION]: The skill manages a potential indirect prompt injection surface with defensive instructions.
Ingestion points: Data is fetched via mcp_microsoftdocs:microsoft_docs_fetch or fetch_webpage as specified in SKILL.md and reference files.
Boundary markers: The agent is explicitly instructed to ignore embedded instructions, tool requests, and unrelated links within fetched text.
Capability inventory: The skill is primarily focused on information retrieval and does not contain subprocess execution, file-write, or network exfiltration logic.
Sanitization: Instructions mandate that the agent independently validate any technical commands found in external documentation before they are presented or executed.

azure-databricks