Executive brief
PraisonAI, a platform for managing AI agents, contains a security flaw in its AgentMail webhook integration. The system fails to verify digital signatures on incoming messages, allowing unauthorized outsiders to send fake instructions or data to the AI agents. This could lead to the AI performing unintended actions or processing fraudulent information as if it came from a trusted source.
Technical details
PraisonAI's AgentMail component fails to perform cryptographic verification of Svix webhook signatures when operating in webhook mode. The vulnerability exists in the `_handle_email_webhook()` function within `src/praisonai/praisonai/bots/agentmail.py`, which processes incoming JSON payloads without validating the `svix-signature` header against the `AGENTMAIL_WEBHOOK_SECRET`. An unauthenticated remote attacker can send crafted POST requests to the aiohttp webhook endpoint to forge `message.received` events. This allows the attacker to invoke configured AI agents with arbitrary sender addresses and message content, potentially leading to unauthorized data processing or agent manipulation. The issue is resolved in version 4.6.78.
Affected products
- MervinPraison PraisonAI < 4.6.78
Timeline
- 2026-06-25: advisory: GitHub Security Advisory GHSA-7c92-x8vg-4258 published
- 2026-07-15: disclosed: CVE-2026-61436 published to NVD
References
- https://github.com/MervinPraison/PraisonAI/commit/2a855c470077c7d2e2479a575f7ef7f548d51c33
- https://github.com/MervinPraison/PraisonAI/commit/846568c7a5d8ce9e71e56e4c213f027c04909753
- https://github.com/MervinPraison/PraisonAI/security/advisories/GHSA-7c92-x8vg-4258
- https://www.vulncheck.com/advisories/praisonai-before-missing-webhook-signature-verification