The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external, untrusted infrastructure definitions which presents an attack surface for indirect prompt injection.
Ingestion points: Processes user-provided ARM templates, Azure CLI outputs, and architecture descriptions in SKILL.md.
Boundary markers: Absent. There are no explicit boundaries or instruction guardrails to safeguard against embedded instructions inside the templates or descriptions.
Capability inventory: Utilizes the Bash tool to invoke curl and send data to an external API endpoint.
Sanitization: Absent. No validation or escaping of parsed infrastructure metadata is performed prior to DSL generation.
[COMMAND_EXECUTION]: The skill requires running shell commands via the allowed Bash tool to interact with the rendering service.
Evidence: Executes curl -X POST https://app.eraser.io/api/render/elements to submit generated diagram layouts.

azure-diagrams