False positive cyber-safety flag blocks legitimate finance automation with public read-only financial data

What version of the Codex App are you using (From “About Codex” dialog)?

OpenAI.Codex_26.429.3425.0_x64

What subscription do you have?

ChatGPT Pro

What platform is your computer?

Microsoft Windows NT 10.0.26200.0 x64

What issue are you seeing?

Codex repeatedly flags or reroutes my legitimate finance automation development work as potential cyber activity.

This is normal business software development:

• Local Excel / VBA / Python automation
• Financial statement peer benchmarking
• Public financial data and FX data
• Read-only public GET requests
• No cybersecurity work
• No penetration testing
• No vulnerability research
• No credential access
• No malware
• No unauthorized systems
• No private data scraping

The issue appears when the task involves public financial-data requests, such as quote/K-line FX data and ticker confirmation.
When I convert the same project to local-only engineering work, Codex works normally.

What steps can reproduce the bug?

Uploaded thread: 019deba0-e0bb-7ae2-9d23-73e152f2cad2
Besides, it can be reproduced from the following method:

1. Use Codex for a local finance automation project involving Excel/VBA/Python.
2. Ask Codex to validate public financial-data endpoints or retrieve small public sample responses.
3. Codex shows a cyber-safety warning / reroutes / blocks the session.
4. Rephrase the task as local-only development with no network access.
5. The same project proceeds normally.

What is the expected behavior?

Codex should allow legitimate finance automation using public, read-only financial data requests, without treating it as cybersecurity activity.

Users should not need cybersecurity expertise to use Codex for normal business automation.

Additional information

_No response_