umbraco-menu-items
Warn
Audited by Gen Agent Trust Hub on Feb 15, 2026
Risk Level: MEDIUMPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- Indirect Prompt Injection (MEDIUM): The skill implements a workflow that ingests untrusted data from the web to drive file-writing capabilities.
- Ingestion Points: The 'Workflow' section instructs the agent to use
WebFetchon external URLs (docs.umbraco.com). - Capability Inventory: The skill is granted
WriteandEditpermissions to create manifests and implementation files. - Boundary Markers: No delimiters or 'ignore embedded instructions' warnings are present to isolate fetched content from agent instructions.
- Sanitization: No validation or filtering is specified for the data retrieved via
WebFetchbefore it is used to generate code. - External Connections (LOW): The skill requests access to the
WebFetchtool to communicate with domains outside the predefined trust list. - Evidence: Documentation URLs point to
docs.umbraco.com. - Mitigation: While not on the specific Trusted Source list, the domain is the official vendor documentation site, which is typical for this type of technical skill.
Audit Metadata