The Agent Skills Directory

Build Command Execution: The skill has the capability to run package manager commands such as npm install, pip install, and dotnet build to verify application health. To minimize risk, the skill enforces an absolute prohibition on these commands during standard evaluation, limiting them only to scenarios where the agent has authored or remediated the code. All such executions are strictly gated by an ask_user requirement, ensuring that no commands are run without explicit per-command consent from the user.
Indirect Prompt Injection Surface: The skill functions by analyzing user-provided source code and project manifests to determine deployment requirements. Because the skill ingests untrusted content from the repository, it is inherently susceptible to indirect prompt injection where malicious instructions within the code could attempt to manipulate the agent. The skill mitigates this by favoring static analysis and explicitly prohibiting the execution of untrusted scripts or test suites found in the repository.
Azure Environment Integration: The skill interacts with the Azure CLI to retrieve subscription details and user identity for deployment context. This access is necessary for the skill's primary purpose of preparing an application for Azure and is handled using standard CLI tools provided by the vendor.

azure-app-onboard-prereq