A report ties Claude Code's 'Don't Ask Again' option to 2,539 passwords and access keys exposed on GitHub. Here is what to check if your team uses AI coding tools.
A report from OpenTools says 2,539 secrets, such as passwords and access keys, were exposed on GitHub. It ties the problem to an option in Claude Code called "Don't Ask Again." Claude Code is an AI coding tool made by Anthropic.
GitHub is a site where developers store and share code. If a password or access key ends up in that code, anyone who finds it may be able to use it. An access key works like a password that lets one program connect to another service. Those are the "secrets" in this story.
Claude Code helps people write software with the help of AI. "Don't Ask Again" is one of its options. The report links that option to the exposed secrets.
If you or your developers use AI coding tools, review what you have approved. Check any saved choices, including options that stop the tool from asking for permission. Keep passwords and access keys out of code. They belong in a separate, protected place, not in files that get shared or uploaded.
If your team uses AI coding tools, review the permissions you have approved and keep passwords and access keys out of your code.
Goldman Sachs has limited AI coding tools to its engineers, a sign that software built by non-engineers near client data and money is getting a harder look.
Read more: Goldman Sachs limits AI coding tools to its engineersOpenAI plans to buy Pierre Computer Company, whose code review tools already run inside Codex, a sign that AI coding is shifting toward checking code, not just writing it.
Read more: OpenAI Is Buying a Small Company That Helps Teams Review CodeA consultant at AlixPartners says Claude Code overlooked $800,000 in one merger contract. Anyone using AI to read deals or invoices should take note.
Read more: Consultant says AI tool missed $800,000 in a single deal contract
