codex - 💡(How to fix) Fix Model permission UI shows Full Access, but runtime remains restricted with no escalation path [7 comments, 7 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
openai/codex#19356Fetched 2026-04-25 06:10:56
View on GitHub
Comments
7
Participants
7
Timeline
17
Reactions
0
Timeline (top)
commented ×7labeled ×5cross-referenced ×3unlabeled ×2

Error Message

Error: listen EPERM: operation not permitted ::1:5173

Root Cause

The agent also could not request escalation because approval_policy was never.

RAW_BUFFERClick to expand / collapse

What version of Codex CLI is running?

codex-cli 0.124.0

What subscription do you have?

Pro

Which model were you using?

gpt-5.5

What platform is your computer?

Darwin 25.3.0 arm64 arm

What terminal emulator and version are you using (if applicable)?

iTerm2

What issue are you seeing?

Title: Model permission UI shows Full Access, but runtime remains restricted with no escalation path

Observed: The UI showed: 3. Full Access (current) - Codex can edit files outside this workspace and access the internet without asking.

But the model/runtime context still reported: sandbox_mode = workspace-write network access = restricted approval policy = never

Impact: The agent could not bind a local Vite dev server socket or kill/restart a local process. npm run dev failed with: Error: listen EPERM: operation not permitted ::1:5173

The agent also could not request escalation because approval_policy was never.

After switching to: 2. Auto-review

The runtime exposed an escalation path, and the same dev-server action succeeded via require_escalated. Vite started at http://localhost:5174/.

Expected: Full Access should either actually provide unrestricted execution, or the runtime context should match the UI state. It should not be less capable than Auto-review.

Actual: Full Access appeared selected in UI, but the runtime behaved like restricted/no-approval mode.

Likely area: Permission state propagation between UI selection and model tool runtime, possibly after model/session switch.

What steps can reproduce the bug?

Re-run the run dev

What is the expected behavior?

3 full access works.

Additional information

No response

extent analysis

TL;DR

Switching to "Auto-review" mode may provide a temporary workaround for the issue where "Full Access" mode does not grant unrestricted execution as expected.

Guidance

  • Verify that the issue persists after switching to "Auto-review" mode and using the require_escalated option to confirm if the problem is specific to the "Full Access" mode.
  • Check if the model and session switch is properly handled in the permission state propagation between the UI selection and the model tool runtime.
  • Test if the issue is specific to the gpt-5.5 model or if it occurs with other models as well.
  • Review the approval policy settings to ensure that it is not overriding the "Full Access" mode.

Example

No code snippet is provided as the issue does not contain specific code references.

Notes

The issue seems to be related to the permission state propagation between the UI selection and the model tool runtime, possibly after a model or session switch. The provided information does not contain enough details to pinpoint the exact cause.

Recommendation

Apply workaround: Switch to "Auto-review" mode until the issue with "Full Access" mode is resolved, as it provides an escalation path and allows for unrestricted execution.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING

codex - 💡(How to fix) Fix Model permission UI shows Full Access, but runtime remains restricted with no escalation path [7 comments, 7 participants]