kelly-behavior-predict
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill operates as an analytics dashboard that ingests potentially untrusted data from external Busabase bases, creating a surface for indirect prompt injection.
- Ingestion points: Data is ingested from the 'sessions', 'segments', and 'settings' bases within
content/kelly-behavior-predict-app/app/js/providers/busabase-provider.js. - Boundary markers: The application uses HTML escaping to delimit text in the user interface, though it lacks explicit prompt boundary markers for the AI agent when reading the processed data.
- Capability inventory: The skill is restricted to standard
busabase-sdkoperations for record management and does not possess capabilities for arbitrary command execution or non-whitelisted network access at runtime. - Sanitization: Data rendered in the AirApp dashboard is sanitized using HTML entity escaping via the
escapeHtmlfunction inapp.js.
Audit Metadata