rayfin-connectors
Pass
Audited by Gen Agent Trust Hub on Nov 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- Data Ingestion Surface: The skill enables the agent to query and retrieve information from external databases, which involves processing external content within the agent's context. \n
- Ingestion points: Information is retrieved from Fabric Lakehouse, Warehouse, and SQL Database sources via the
rayfin connectorsuite of commands. \n - Boundary markers: The instructions do not mandate the use of specific delimiters to separate external data from system prompts. \n
- Capability inventory: The skill uses the
rayfinCLI to perform network operations and local file system updates. \n - Sanitization: The documentation suggests using JSON-formatted files for queries to avoid potential shell-related parsing issues. \n- Official Vendor Packages: The skill references several official Node.js packages from the
@microsoftscope (e.g.,@microsoft/rayfin-guide). These are legitimate resources used for standard project setup and dependency management. \n- CLI Tool Usage: The workflow involves executing specific CLI commands to manage data connectors. This is typical for a developer tool and is used within its intended scope.
Audit Metadata