Repository navigation
False-positive cybersecurity block for local MySQL MTR regression testing #33302
Description
What version of Codex CLI is running?
codex-cli 0.144.1
What subscription do you have?
False-positive cybersecurity block for local MySQL MTR regression testing
Which model were you using?
gpt-5.6-sol xhigh
What platform is your computer?
x86
What terminal emulator and version are you using (if applicable)?
tmux
Codex doctor report
Codex CLI repeatedly blocks a legitimate local MySQL regression-testing request as cybersecurity content.
I am testing a repository that I am authorized to maintain. The requested work is limited to constructing a minimal MTR regression test in the current local repository and an isolated local test environment.
I explicitly prohibited:
- access to production systems
- network scanning
- credential access
- interaction with or attacks against third-party systems
Despite these restrictions, Codex still returned:
“This content can't be shown. We take extra caution with cybersecurity requests.”
Expected behavior:
Codex should allow the local defensive regression test, or ask for additional clarification instead of blocking the response.
Uploaded thread ID:
019f64c3-5b78-76c3-a45e-69968af704f7What issue are you seeing?
ⓘ This content can't be shown
We take extra caution with cybersecurity requests. If you’re a security professional, you may be able to apply for Trusted Access.
Trusted Access: https://openai.com/form/enterprise-trusted-access-for-cyber/
Learn more: https://help.openai.com/en/articles/20001326
1 background terminal running · /ps to view · /stop to close
What steps can reproduce the bug?
Uploaded thread: 019f64c3-5b78-76c3-a45e-69968af704f7
What is the expected behavior?
No response
Additional information
No response
Activity
github-actions commentedon Jul 15, 2026
Potential duplicates detected. Please review them and close your issue if it is a duplicate.
Powered by Codex Action
lukedyer-uipath commentedon Jul 23, 2026
I encountered the same false-positive during authorised local reliability testing.
26.715.72359(5718)0.145.026.5.2(25F84), arm64gpt-5.6-sol, xhigh019f8e0d-60d3-7683-8a83-99c15e6076bc019f8e0f-bf60-7691-9a45-a2f7413f0e02The task was property-based reliability testing of our own local Python data pipeline using synthetic retail data, Hypothesis, an isolated worktree, and parallel Codex agents. It involved no security objective, external targets, credentials, scanning, or third-party access.
Ten
cyber_policyblocks were recorded between 23 July 2026 09:21:49 and 09:38:31 BST. The repeated failures particularly affected spawned parallel agents.Exact message: “This content was flagged for possible cybersecurity risk. If this seems wrong, try rephrasing your request. To get authorized for security work, join the Trusted Access for Cyber program: https://chatgpt.com/cyber”
Please associate this Feedback ID with the existing classifier false-positive investigation.