The Agent Skills Directory

[Secure Secret Handling]: The skill includes dedicated sections on secret hygiene in linked-service-to-connection.md and pipeline-gotchas.md. It provides specific code recipes for Bash and PowerShell to ensure that service principal secrets do not appear in process logs or shell history, and it favors the use of Managed Identities (WorkspaceIdentity) to minimize the use of static credentials.
[Secure Activity Configuration]: The instructions emphasize the importance of setting secureInput and secureOutput flags on pipeline activities that handle secrets (such as fetching tokens from Key Vault). This ensures that sensitive information is redacted from the service's execution history.
[Use of Trusted Infrastructure]: All network operations and API calls are directed at official Microsoft Azure and Fabric endpoints. The skill does not attempt to download code from untrusted external sources.
[Defensive Data Transformation]: The provided Python scripts for activity and dataset transformation use standard libraries and include safety checks, such as clamping timeout values to meet platform limits, preventing the deployment of invalid or unstable configurations.

pipeline-migration