Anchorr, a Discord bot used for media requests, contains a stored cross‑site scripting flaw in the web dashboard’s User Mapping dropdown. An attacker who is an unprivileged member of the configured Discord guild can insert malicious JavaScript into this field. When the bot’s administrator later visits the dashboard, the stored payload is executed in the administrator’s browser, enabling the attacker to run arbitrary code with the same privileges as the admin. The attacker can then call the unprotected GET /api/config endpoint, which returns all credentials in plaintext, including DISCORD_TOKEN, JELLYFIN_API_KEY, JELLYSEERR_API_KEY, JWT_SECRET, WEBHOOK_SECRET, and bcrypt password hashes. The vulnerability is categorized as CWE‑79 and CWE‑200 and carries a CVSS score of 9.7.

The issue affects the openVESSL Anchorr bot version 1.4.1 and all earlier releases. The bug exists in the web dashboard component of the bot. Users running these versions and operating the administrative dashboard are vulnerable. Updates to Anchorr v1.4.2 and later contain the fix.

The flaw is high‑severity with an EPS score below 1 % and is not listed in the CISA KEV catalog, indicating low current exploitation activity. However, the vulnerability can be leveraged without any authentication against Anchorr itself; the only requirement is that the attacker has Discord membership in the guild configured for the bot. Because the stored payload runs in the administrator’s browser, any compromised admin session leads to full credential theft. The attack path is straightforward: an attacker injects malicious JavaScript via the dropdown, the next admin who views the dashboard triggers the payload, and the attacker exfiltrates secrets via the unsecured /api/config endpoint.