The flaw originates from a path confusion in AutoGPT's webhook ingestion route, wherein the provider URL segment is taken at face value without validating the webhook.provider field. The result is that a request to /compass/webhooks/{webhook_id}/ingress can trigger the CompassWebhookManager, whose BaseWebhooksManager.verify_signature is a no‑op. Consequently, the typical X-Webhook-Secret check is bypassed, allowing an attacker to execute any generic webhook graph under the ownership of the configured webhook. This constitutes an authentication bypass (CWE‑287) that effectively enables arbitrary code execution as the webhook owner.

Affected are all releases of Significant‑Gravitas AutoGPT prior to version 0.6.70. The vulnerability resides in autogpt_platform/backend/backend/api/features/integrations/router.py and impacts any instance that exposes the /compass/webhooks/{webhook_id}/ingress endpoint and uses CompassWebhookManager for webhook handling. The 0.6.70 release includes the fix that enforces provider verification.

The CVSS score of 8.2 signals high severity; no EPSS data is available, and the vulnerability is not currently listed in CISA KEV. Attackers can trigger it remotely by sending a crafted HTTP request to the vulnerable endpoint. As the secret verification bypass is straightforward, the risk of exploitation remains high, especially for environments that allow unauthenticated webhook ingestion. Immediate patching is recommended.