OpenClaw has a self-attack vulnerability that mistakenly executes Bash commands, leading to key leakage
Web3 security company GoPlus stated that the AI development tool OpenClaw has recently been reported to have experienced a self-attack security incident. During the execution of automated tasks, the system constructed an incorrect Bash command while calling Shell commands to create a GitHub Issue, inadvertently triggering command injection, which led to the exposure of a large number of sensitive environment variables.
In the incident, the AI-generated string contained a set wrapped in backticks, which was interpreted by Bash as command substitution and executed automatically. Since Bash outputs all current environment variables when executing set without parameters, this ultimately resulted in over 100 lines of sensitive information (including Telegram keys, authentication tokens, etc.) being directly written to the GitHub Issue and publicly published. GoPlus recommends that in AI automation development or testing scenarios, API calls should be used instead of directly concatenating Shell commands, and the principle of least privilege should be followed to isolate environment variables. Additionally, high-risk execution modes should be disabled, and a manual review mechanism should be introduced for critical operations.
链捕手ChainCatcher nhắc nhở: hãy nhìn nhận blockchain một cách tỉnh táo, nâng cao ý thức rủi ro, cảnh giác với việc phát hành và đầu cơ các token ảo. Mọi nội dung trên trang chỉ là thông tin thị trường hoặc quan điểm của các bên liên quan, không cấu thành bất kỳ hình thức tư vấn đầu tư nào. Nếu phát hiện nội dung nhạy cảm, hãy nhấp , chúng tôi sẽ xử lý kịp thời.






