kelly-messenger
Pass
Audited by Gen Agent Trust Hub on Aug 25, 2026
Risk Level: SAFE
Full Analysis
- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill interfaces with well-known messaging services (Slack, Discord, Telegram, and WhatsApp/Facebook Graph API) and a user-provided Busabase instance. Authentication tokens for these services are managed exclusively through local environment variables and are not stored in the database or committed to version control.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external chat data from platforms like WhatsApp, Slack, Discord, and Telegram through its sync and ingestion scripts (scripts/sync_messages.mjs, scripts/ingest_messages.mjs). This ingestion surface is mitigated by the application's core design, which requires explicit human review and approval in the Outbox queue before any generated reply is sent. Technical sanitization includes the use of escapeHtml throughout the UI rendering components (app/app/app.js, app/app/js/message-views.js).
- [COMMAND_EXECUTION]: The skill utilizes deterministic Node.js scripts for management tasks. These scripts (scripts/sync_messages.mjs, scripts/send_outbox.mjs) perform API calls to messaging platforms and CRUD operations on the user's Busabase backend, operating with appropriate permissions scoped to the user's configured accounts.
Audit Metadata