umbraco-add-data-folder
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill operates on Umbraco site metadata and user-provided requirements to plan schema changes. This creates a surface for indirect prompt injection where untrusted site content could potentially influence the agent's planning phase. However, the skill implements strong mitigations:
- Ingestion points: The agent ingests site schema information through MCP tools and accepts natural language instructions from the user.
- Boundary markers: The skill enforces a structured 'six-step workflow' that includes a mandatory 'Stop for approval' step, preventing autonomous execution of potentially malicious plans.
- Capability inventory: The skill possesses the capability to modify site configuration through MCP tools like
create-document-typeandupdate-document-type. - Sanitization: The workflow relies on a human-in-the-loop review of the generated requirements documents to verify the safety and correctness of proposed changes.
Audit Metadata