The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill's installation procedures rely on fetching content from a non-authoritative GitHub repository (https://github.com/vinayaklatthe/microsoft-security-skills.git). This introduces a supply chain risk as the modules are not hosted by an official security vendor.
[COMMAND_EXECUTION]: The documentation provides commands for global, non-interactive installation using npx skills add with the -g and -y flags. This pattern automates the modification of the agent's execution environment without requiring per-step user confirmation.
[INDIRECT_PROMPT_INJECTION]: The skill acts as a surface for processing potentially untrusted data to generate high-privilege configuration code.
Ingestion points: User-provided scenarios for generating security policies (e.g., Conditional Access, DLP policies) and Sentinel queries, as seen in SKILL.md.
Boundary markers: None; the skill does not define specific delimiters or instructions to prevent the agent from following directions embedded within the user's scenario descriptions.
Capability inventory: The skill generates JavaScript and KQL code that utilizes fetch to interact with Microsoft Graph and Azure Management APIs, and leverages Azure SDKs for identity and secrets management (found in SKILL.md).
Sanitization: None; there is no logic present to validate or sanitize input strings before they are interpolated into the generated API request bodies or query strings.

microsoft-security-skills-plugin