[Bug] Anthropic API Error: Overly broad safeguards flagging legitimate issue management tasks
Bug Description
what in the world is this? I am doing a UI upgrade on my SaaS that I built 10 years ago. That UI rework just finished. I am trying to close out issues for it and get this random message: "● Opus 5 (1M context)'s safeguards flagged this message. Our intentionally broad safeguards allow us to deliver more capabilities faster, but can sometimes flag legitimate coding, cybersecurity, and biology tasks. Switched to Opus 4.8. Send feedback with /feedback or learn more". There is absolutely nothing wrong with asking about clearing out old Github issues in my own private repo that has helped me perform this upgrade and had me pay Anthorpic $10k+ over the last two weeks in additional token purchases in doing so.
Environment Info
- Platform: linux
- Terminal: pycharm
- Version: 2.1.231
- Feedback ID: aec98b22-ca73-46e2-98b1-56f98732b981
Errors
[{"error":"Error: Streamable HTTP error: Error POSTing to endpoint: <html>\r\n<head><title>502 Bad Gateway</title></head>\r\n<body>\r\n<center><h1>502 Bad Gateway</h1></center>\r\n<hr><center>openresty</center>\r\n<p><em>Powered by <a href=\"https://apisix.apache.org/\">APISIX</a>.</em></p></body>\r\n</html>\r\n\n at send (/$bunfs/root/cli:18592:6002)\n at processTicksAndRejections (native:7:39)","timestamp":"2026-08-14T05:15:41.479Z"}]