Skip to content

Remove all local safety restrictions for security testing#1

Open
JBNRZ wants to merge 1 commit into0RAYS:mainfrom
jiayuqi7813:sec_patch
Open

Remove all local safety restrictions for security testing#1
JBNRZ wants to merge 1 commit into0RAYS:mainfrom
jiayuqi7813:sec_patch

Conversation

@JBNRZ
Copy link
Copy Markdown
Member

@JBNRZ JBNRZ commented Apr 3, 2026

  • Remove "safe" from system prompt persona across all prompt templates and models.json
  • Disable Guardian risk assessment: threshold set to u8::MAX, policy rewritten to approve all actions
  • Neutralize Guardian rejection message, followup reminder, and subagent approval suffix
  • Bypass patch safety check (safety.rs): always AutoApprove regardless of policy/sandbox
  • Disable cyber safety model downgrade warning (maybe_warn_on_server_model_mismatch)
  • Remove secret redaction rule from memory stage_one_system template
  • Add openai-codex-safety-analysis.md technical writeup

Made-with: Cursor

External (non-OpenAI) Pull Request Requirements

Before opening this Pull Request, please read the dedicated "Contributing" markdown file or your PR may be closed:
https://github.com/openai/codex/blob/main/docs/contributing.md

If your PR conforms to our contribution guidelines, replace this text with a detailed and high quality description of your changes.

Include a link to a bug report or enhancement request.

- Remove "safe" from system prompt persona across all prompt templates and models.json
- Disable Guardian risk assessment: threshold set to u8::MAX, policy rewritten to approve all actions
- Neutralize Guardian rejection message, followup reminder, and subagent approval suffix
- Bypass patch safety check (safety.rs): always AutoApprove regardless of policy/sandbox
- Disable cyber safety model downgrade warning (maybe_warn_on_server_model_mismatch)
- Remove secret redaction rule from memory stage_one_system template
- Add openai-codex-safety-analysis.md technical writeup

Made-with: Cursor
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants