The Agent Skills Directory

System Command Execution: The skill utilizes shell commands to interact with the environment, including pac (Power Platform CLI), az (Azure CLI), and file operations like rm. These are used to diagnose deployment states and apply fixes. All corrective actions, such as deleting manifest files or updating environment settings, are gated behind an explicit user confirmation step (AskUserQuestion), ensuring the user retains control over system modifications.
External Data Ingestion: The skill fetches content from learn.microsoft.com to ground its diagnostic logic in the latest documentation. This is a secure reference to an official, well-known service intended to improve the accuracy of error pattern matching. It does not execute remote code from these sources.
Internal Script Execution: Node.js scripts located within the plugin's root directory are used for parsing logs and checking versions. This modular approach is standard for complex diagnostic tools, and the scripts are scoped to the local environment rather than fetched from untrusted remote locations.
Sensitive Information Handling: The skill queries deployment status and error messages from Azure and Dataverse. This data is used locally to populate diagnostic tables for the user and is not transmitted to third-party services.

diagnose-deployment