fcode-ui-triggers
Pass
Audited by Gen Agent Trust Hub on Aug 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill describes an architecture for ingesting data into automated processes through UI buttons and forms.
- Ingestion points: Processes receive data from
locationIdcontext parameters and user-submitted forms described inSKILL.md. - Boundary markers: The documentation explicitly distinguishes between server-side injected (trusted) parameters and client-editable (untrusted) form parameters.
- Capability inventory: The skill enables the execution of processes that can return data to the UI or perform background actions via
fcode.processes.run. - Sanitization: The skill provides clear security guidance for developers to avoid authorization vulnerabilities: "Never authorize on company_id / triggered_from_location in a form... they arrive as pre-filled, client-editable fields."
Audit Metadata