automate-whatsapp
Pass
Audited by Gen Agent Trust Hub on Mar 11, 2026
Risk Level: SAFEPROMPT_INJECTIONREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill facilitates the creation of agents that ingest untrusted user input from WhatsApp, such as in 'assets/workflow-linear.json' (vars.user_name) and 'assets/workflow-customer-support-intake-agent.json' (vars.support_intake). These ingestion points create a surface for indirect prompt injection. Capability Inventory: The skill performs network operations via fetch in 'scripts/lib/workflows/kapso-api.js', 'scripts/lib/databases/kapso-api.js', and 'scripts/lib/functions/kapso-api.js' to manage remote functions, triggers, and database rows. Boundary Markers: No explicit boundary markers or 'ignore' instructions are present in the provided workflow assets. Sanitization: No sanitization of ingested WhatsApp message content is performed before processing by agent nodes.- [REMOTE_CODE_EXECUTION]: The skill includes functionality to manage and deploy JavaScript code to a remote runtime (Kapso) through scripts like 'create-function.js' and 'deploy-function.js'. This is a primary intended use-case for building custom automation logic.- [EXTERNAL_DOWNLOADS]: The 'scripts/openapi-explore.mjs' utility fetches OpenAPI documentation from 'docs.kapso.ai'. This is a legitimate operation for API discovery and document lookup for the vendor's platform.
Audit Metadata