A newly filed bug report on GitHub details a potential session or cache leakage vulnerability in Claude Code, Anthropic's developer-focused AI tool. The issue, posted to the claude-code repository, specifically warns of data crossing between workspace instances or consumer accounts.
The report has drawn attention on Hacker News, where it gained 14 points and three comments within an hour of publication. The timing is sensitive, coming as AI-assisted coding tools face increasing scrutiny over data isolation and security practices.
At this stage, the claim rests solely on the GitHub issue; no official confirmation or denial has been issued by Anthropic. The repository's maintainers have not yet labeled the report or responded publicly to the allegations.
If validated, the bug could have implications for enterprises using Claude Code across multiple projects, potentially exposing sensitive code or internal data between workspaces. Developers and security teams may want to monitor upstream patches closely.
Anthropic has yet to comment on the report's veracity or timeline for a fix.