The vulnerability is an insecure direct object reference, allowing a remote attacker to view private chatbot conversations of other users by altering the conversation ID in the request URL. This results in disclosure of potentially sensitive or confidential user data without requiring authentication or impersonation. The weakness, identified as CWE‑639, indicates that data ownership and access control checks are insufficient, exposing user‑generated content directly.

Affected products are 1millionbot Millie chat. All releases prior to the patched 3.6.0 version contain the IDOR flaw. The exact affected version range is not specified, but the vendor confirmed that 3.6.0 includes the fix.

The CVSS score of 7 indicates high severity, and the EPSS score is below 1%, suggesting low exploitation probability; the vulnerability is not listed in the KEV catalog. The bug does not require credentials; an attacker merely needs a valid conversation identifier. If an attacker can obtain or enumerate such IDs, they can retrieve private conversations. Given the simple request structure, the attack vector is remote and straightforward, making potential exploitation feasible in realistic scenarios.