The Agent Skills Directory

Command Execution: The skill utilizes various command-line tools such as az (Azure CLI), pac (Power Platform CLI), and git. It also executes multiple local helper scripts using node to automate version checks, validate environment prerequisites, and manage Dataverse environment variables. These tools are used to acquire access tokens and synchronize project state with the remote environment.
Indirect Prompt Injection Surface: The skill processes data from external project files (such as YAML site settings and solution manifests) and Dataverse API responses to drive its logic. This data ingestion represents a potential surface where malformed or untrusted external data could influence the agent's actions.
Ingestion points: Data is read from .solution-manifest.json, .powerpages-site/website.yml, various site setting YAML files, and Dataverse OData API responses.
Boundary markers: The skill does not explicitly define delimiters for untrusted data during interpolation into internal logic.
Capability inventory: The skill possesses the ability to write project files (deployment-settings.json), execute local shell scripts, perform Git commits, and make network requests to Dataverse environments.
Sanitization: The skill includes specialized validation logic, such as validate-deployment-settings.js, which checks that secret references follow canonical Azure Key Vault URI formats before they are persisted to the project configuration.

configure-env-variables