OpenClaw logs raw Telegram bot tokens in error messages and stack traces when request URLs contain the bot token. The exposed token enables an attacker to impersonate the bot and gain full Bot API access. This is a credential exposure weakness, classified as CWE‑522, and could compromise the confidentiality and integrity of bot-controlled interactions.

OpenClaw, version 2026.2.14 and earlier on the OpenClaw personal AI assistant platform, runs on node.js. Any user deploying or running this application before 2026.2.15 may have written bot tokens into logs, crash reports, CI output, or support bundles.

The CVSS score of 6.9 indicates moderate severity, while the EPSS score of less than 1% suggests a low likelihood of exploitation at this time. The vulnerability is not currently cataloged in CISA's KEV list. Attackers would need access to application logs or other outputs to retrieve the compromised token, implying a local or compromised system or compromised CI pipeline as the probable attack vector. Once the token is obtained, the attacker can use the Telegram Bot API to impersonate the bot, send messages, and perform any action allowed by the bot's permissions.