fabriciq-ontology-authoring-cli
Pass
Audited by Gen Agent Trust Hub on Jul 10, 2026
Risk Level: SAFE
Full Analysis
- Administrative Tool Integration: The skill uses standard command-line tools including
az rest,curl,jq, andbase64to interact with the Microsoft Fabric control plane. These operations are performed against official Microsoft domains (api.fabric.microsoft.com) and are necessary for the skill's core functionality. - Data Schema Ingestion: The skill processes table schemas from Fabric Lakehouse and Eventhouse environments to facilitate data bindings. It includes specific guidance on discovering these schemas using internal consumption skills or standard SQL/KQL queries, which is appropriate for the authoring use case.
- Indirect Prompt Injection Guardrails: A robust 'Preview & Confirm' mechanism is implemented as a security boundary. Before any modifications are made via
createItemorupdateDefinition, the agent is required to display an ASCII-based visualization or a change-set diff to the user. This ensures that the user can verify the data being used to build the ontology and provides a manual checkpoint to prevent unintended modifications. - Secure Version Maintenance: The skill includes an 'Update Check' routine that uses
gitto compare local metadata against the remote repository. This allows the agent to inform the user of newer versions, ensuring that the skill remains up-to-date with current safety guidelines and features. - Asynchronous Operation Management: The skill provides detailed patterns for handling Fabric's Long-Running Operations (LROs). It correctly captures response headers and polls status endpoints on the primary Fabric host, ensuring that administrative tasks are completed reliably and that any API errors are accurately reported.
Audit Metadata