kelly-insure-data
Pass
Audited by Gen Agent Trust Hub on Aug 25, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill includes maintenance scripts such as
scripts/backfill_pdf_metadata.mjsthat utilizeexecFileto run system utilities likepdftotextandpdfinfofor processing local PDF files. - [EXTERNAL_DOWNLOADS]: The project defines several Node.js dependencies in its
package.jsonfiles, includinghonoandbusabase-sdk, which are required for the application's runtime and server functionality. - [PROMPT_INJECTION]: The skill processes external data from Busabase, which represents a surface for indirect prompt injection.
- Ingestion points: Insurance records and file metadata are retrieved from the Busabase API in
content/kelly-insure-data-app/app/js/providers/busabase-provider.js. - Boundary markers: The skill does not explicitly define instructions or markers for the AI agent to differentiate between system instructions and the ingested data content.
- Capability inventory: The skill provides an interactive UI for data review, and associated scripts can perform network requests and file operations.
- Sanitization: Content rendered in the user interface is escaped to prevent execution of malicious scripts via
escapeHtmlincontent/kelly-insure-data-app/app/app.js. - [SAFE]: The application is designed as a read-only viewer, ensuring that the primary user interface cannot mutate the canonical dataset. Sensitive information such as API keys is managed through environment variables rather than being hardcoded in the source code.
Audit Metadata