umbraco-current-user-action
Fail
Audited by Gen Agent Trust Hub on Feb 15, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [Indirect Prompt Injection] (HIGH): The skill demonstrates a high-risk capability tier where external data is ingested to drive code generation tasks.
- Ingestion points: The skill explicitly instructs the agent to use
WebFetchon external URLs (docs.umbraco.com) in the 'Workflow' section. - Boundary markers: Absent. There are no instructions to delimit or treat the external documentation as untrusted data.
- Capability inventory: The skill uses
WriteandEdittools to create manifest files and action implementation classes (.tsand.jsfiles). - Sanitization: None. The agent is directed to generate complete working code based directly on the latest fetched docs without any validation or sanitization steps.
- [External Downloads] (LOW): The skill utilizes
WebFetchto retrieve information fromdocs.umbraco.com. - Evidence: URLs provided in the 'Documentation' section for extension types and registries.
- Trust Scope: This finding is downgraded to LOW per [TRUST-SCOPE-RULE] as the source is the official documentation for the targeted framework. However, the downstream behavior of generating code from this data remains a HIGH severity risk.
Recommendations
- AI detected serious security threats
Audit Metadata