vapi-demo
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill guides the agent to use the Railway CLI, Node.js, and standard Unix utilities (curl, jq) for managing deployments and interacting with the Vapi API.
- [EXTERNAL_DOWNLOADS]: Dependencies like the Vapi Web SDK are retrieved from established services like esm.sh, and the skill provides setup instructions for standard development tools such as Playwright and the Railway CLI.
- [INDIRECT_PROMPT_INJECTION]: The dashboard ingests live transcript and tool call data from Vapi webhooks; the implementation consistently uses
textContentfor DOM updates, which prevents the execution of malicious scripts embedded in the transcript data. - [SAFE]: Secret management follows established best practices, including instructions to use .env files with corresponding .gitignore entries and the use of Railway's environment variables for production secrets.
Audit Metadata