whapi

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 1.00). The skill explicitly instructs agents to ingest and process incoming WhatsApp messages (user-generated content) delivered via webhooks—see references/pattern-bot.md and references/recv-webhooks.md which show extracting messages[n].text.body and passing it to the AI—so untrusted third‑party input can directly influence subsequent tool calls and behavior.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill requires running "npx -y whapi-mcp@latest" (fetching and executing the remote npm package whapi-mcp at runtime), which downloads and runs remote code the agent depends on to provide MCP tools.