openai / openai/codex

Luna Reserve elevated commands are always auto-denied

Open
#43,890 2 comments 2 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

app bug rate-limits sandbox windows-os
Dominant language
Rust
Stars
125k
Forks
19.4k
PR merge metrics
PR metrics pending

Description

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

26.901.51231

What subscription do you have?

Plus

What platform is your computer?

Microsoft Windows NT 10.0.26200.0 x64

What issue are you seeing?

On Windows Desktop, after the account entered Luna Reserve, normal sandboxed commands continued to work, but an explicitly authorized elevated dotnet build --no-incremental was rejected before execution with Automatic approval review failed: You've hit your usage limit. The failure appears to affect the auto-review/elevation path specifically, not ordinary Luna execution.

What steps can reproduce the bug?

Run out of 5h window usage (probably the same for weekly limit). Ask the agent anything that requires elevation and watch it get auto-denied

What is the expected behavior?

It can continue to run the validator in Approve for me mode when using Luna Reserve (GPT-Reserve)

Additional information

Also mentioned it here but I think it's actually a distinct issue https://github.com/openai/codex/issues/29717#issuecomment-5551420182
Out of usage so can't investigate myself any further.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start with the Windows Desktop auto-review and elevation path, reproducing the issue after the 5-hour or weekly usage limit is reached. Compare ordinary sandboxed commands with an explicitly authorized elevated dotnet build --no-incremental in Luna Reserve; done means elevated commands are no longer automatically denied solely because the normal usage limit was reached.

Written by the indexing model from the issue text.

Assessment

Tech stack
rust
Domain
authorization, desktop
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Active
Clarity
Mostly clear
Newbie friendliness
48/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.