GitBook Discloses AI Agent Cyberattack and Credential Exposure
GitBook reported a targeted cyberattack involving AI agents that exposed infrastructure credentials and API tokens.
GitBook disclosed a targeted cyberattack on November 2 involving AI agents that granted an attacker unauthorized access to some infrastructure credentials. The intrusion occurred at 7 p.m. Central European Time. GitBook detected the event immediately and moved to contain it. Potentially exposed information includes API tokens, integration secrets, and questions submitted to the company's AI assistant. The company stated it found no evidence that hosted content was modified.
GitBook revoked personal API tokens and invalidated every customer's API tokens on its platform. It instructed customers to rotate existing GitBook API tokens and replace connected credentials. Users of the Model Context Protocol server must authenticate again. GitBook blocked the attacker, addressed the vulnerabilities, and replaced its own infrastructure credentials.
Crypto exchange PancakeSwap temporarily took its documentation offline before moving it away from GitBook. PancakeSwap stated its documentation is not compromised and its app is not affected. The exchange announced its documentation was back online after being moved to infrastructure under its own control. By Tuesday afternoon, GitBook said its application, API, and published documentation were accessible again.
Sources
- Hacker News: front pageGitbook Breached ↗
- DecryptCrypto Docs Host GitBook Hit by AI Agent Security Breach, Urges Users to Rotate Keys ↗
Every sentence of this story was checked against facts quoted from these articles. How we make the news
Spotted an error? Tell us
