DotShield – stop Claude Code and Cursor from reading your .env (github.com)

🤖 AI Summary
DotShield has launched its public beta, a tool designed to protect developers' sensitive information such as API keys and tokens from AI coding assistants like Claude Code and Cursor. The software prevents these agents from reading or printing secret files, and provides an offline scanner that identifies existing secrets in codebases, git histories, and agent transcripts. Key features include the ability to redact secrets from transcripts, add ignoring rules to prevent accidental exposure, and enforce pre-commit checks to block the inclusion of credentials in git commits. This release is significant for the AI/ML community as it addresses growing concerns regarding data privacy and security in the development process, particularly as AI tools become increasingly integrated into workflows. DotShield operates without any network dependencies, ensuring that all scanning and guarding activities remain local to the developer’s machine, thus enhancing security. The tool also includes detailed commands for scanning, fixing potential leaks, and setting up a protective guard, which can operate in a watch mode to allow for initial assessments before enforcement. Developers are encouraged to provide feedback on the beta version to improve its functionality and address any unforeseen issues.
Loading comments...
loading comments...