analytics-dashboard
Pass
Audited by Gen Agent Trust Hub on Aug 3, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security issues were identified. The skill performs data processing and visualization within the agent's safe environment.\n- [PROMPT_INJECTION]: The skill uses the marker 'CRITICAL' in a header to define its primary execution flow ('Auto-start on load'). This is interpreted as a benign instructional directive for state management rather than an attempt to bypass system safety constraints or security filters.\n- [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection as it processes data from user-uploaded Excel files. Ingestion point: LinkedIn Analytics export file (xlsx) in Step 2; Boundary markers: None; Capability inventory: Generates React artifacts and written analysis with no access to shell, filesystem writes, or network operations; Sanitization: None. Given the limited capability tier, this surface represents a standard functional requirement without significant risk of exploitation.
Audit Metadata