The Agent Skills Directory

[Indirect Prompt Injection Surface]: The skill processes user-supplied incident descriptions to generate diagnostic queries and search documentation. While necessary for its function, this represents a surface where external input enters the agent's context.
Ingestion points: The incident-description argument in SKILL.md.
Boundary markers: The skill includes a CAUTION block advising human review of all outputs, though specific technical delimiters for input are not defined.
Capability inventory: The skill can read local project documentation (runbooks/, docs/) and utilize Azure-specific MCP tools as described in SKILL.md.
Sanitization: Not explicitly implemented within the skill instructions; instead, the skill relies on mandatory human review as a safeguard.
[Authorized Resource Access]: The skill is designed to interact with official Microsoft and Azure tools and documentation. This is consistent with its role as an incident response assistant and uses resources provided by the skill's author.

incident-response