Anyone else's OpenAI and Claude API keys mysteriously leaked last night?

As of now, there is no verified evidence of a widespread leak of OpenAI or Claude API keys occurring on a specific "last night" timeframe. Most API key compromises are individual or small-scale, often caused by user error rather than mass platform breaches. If you suspect your API keys have been compromised, here’s what you should do immediately:

Steps to Secure Your API Keys:

1. ·Rotate your keys: Generate new API keys through OpenAI’s API keys dashboard and Claude’s API keys page. Delete the old keys to revoke access.
2. ·Audit usage logs: Check OpenAI’s usage dashboard and Claude’s API logs for unusual activity, such as unrecognized IP addresses or spikes in usage.
3. ·Enable 2FA: Add an extra layer of security to your accounts by enabling two-factor authentication.
4. ·Scan for accidental exposure: Use tools like GitHub’s secret scanning or git-secrets to check for hardcoded keys in your repositories.
5. ·Review third-party integrations: Revoke API access for any untrusted tools and avoid sharing raw API keys with unvetted services.

Common Causes of Key Compromises:

• ·Accidental public exposure: Sharing keys in public repositories or unencrypted channels.
• ·Local machine compromise: Malware or keyloggers on your system.
• ·Phishing attacks: Unauthorized access to your API provider account through phishing.
• ·Misconfigured integrations: Using unsecured third-party tools or improperly managing team access.

Proactive Security Practices:

• ·Rotate keys every 90 days or sooner if shared.
• ·Store keys in environment variables rather than hardcoding them.
• ·Use fine-grained keys to limit permissions to necessary endpoints.

If you find specific evidence of unauthorized activity, such as unexplained charges or API requests, contact OpenAI’s security support or Anthropic’s security team for investigation. Avoid spreading unverified claims of mass leaks, as this can cause unnecessary panic and distract from the user error that’s the root cause of most key compromises.