The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external data from Microsoft Intune (device names, app policies, etc.). While this presents a theoretical attack surface where malicious data could influence agent behavior, the skill includes explicit guardrails for context validation, structured output, and identifier redaction to mitigate risks.
Ingestion points: Microsoft Graph API responses for managed devices and app protection policies.
Boundary markers: Not explicitly defined, though structured output is required.
Capability inventory: The skill uses Bash, Write, and Edit tools to interact with the environment.
Sanitization: Requires redaction of tenant, subscription, and object identifiers before returning output.
[COMMAND_EXECUTION]: The skill has access to the Bash tool to perform Intune operations. The instructions mandate deterministic workflows and fail-fast context validation before any API calls are made, reducing the risk of arbitrary command execution.

Microsoft Intune