Webhook Signature Verification Missing in PraisonAI AgentMail

PraisonAI AgentMail versions prior to 4.6.78 contain a critical authentication bypass vulnerability in their webhook implementation. The vulnerability stems from absent signature verification on incoming webhook events, specifically the message.received event handler. This allows unauthenticated attackers to POST crafted payloads directly to the webhook endpoint, injecting messages with spoofed sender addresses into the agent system. Affected organizations running vulnerable versions of PraisonAI AgentMail are at risk of message injection attacks, where adversaries can manipulate agent responses, trigger automated replies to attacker-controlled addresses, and circumvent configured sender allow/block lists—potentially leading to credential harvesting, social engineering, or further lateral movement within integrated systems.

While this CVE doesn't map to specific MITRE ATT&CK techniques, Casky's platform would identify the underlying attack patterns associated with CWE-290 (Improper Authentication) by analyzing webhook request patterns and authentication enforcement. Practitioners using Casky would observe detection signals around: (1) unauthenticated POST requests to webhook endpoints lacking cryptographic signature validation, (2) message injection attempts with modified sender headers bypassing list-based controls, and (3) anomalous reply routing to external addresses. Casky's extended reasoning capabilities would help practitioners correlate these signals to recognize authentication bypass techniques and implement compensating controls—such as IP allowlisting, request signing mechanisms, or WAF rules—while awaiting patches to version 4.6.78 or later.