Permissions button labels should be verbs, not "Yes" or "No"

Resolved 💬 2 comments Opened Mar 23, 2026 by btraut Closed Jul 1, 2026
💡 Likely answer: A maintainer (etraut-openai, contributor) responded on this thread — see the highlighted reply below.

What issue are you seeing?

Really basic QoL improvement request here:

It's generally good UI design to label buttons with verbs such as "Allow" and "Deny" vs "Yes" or "No". This allows a reader to very quickly scan the prompt and make a more-informed decision without having to read the full title and details.

<img width="1234" height="614" alt="Image" src="https://github.com/user-attachments/assets/e1dea01a-3509-4816-be2c-31485ebb9ddd" />

What steps can reproduce the bug?

Leave permissions on and perform basically any task that requires the agent to leave the sandbox.

What is the expected behavior?

_No response_

Additional information

_No response_

View original on GitHub ↗

2 Comments

Acture · 5 months ago

The relevant code is in codex-rs/tui/src/bottom_pane/approval_overlay.rs (and the mirrored tui_app_server copy).

Prepared a fix (Acture:fix/approval-button-labels-15518):

  • Replace "Yes, ..." → "Allow ..." and "No, ..." → "Deny ..." on all permission approval buttons
  • Mirror changes in both tui and tui_app_server
  • Update all insta snapshots (22 files)
  • Feedback dialog Yes/No left unchanged (different context)

All 1369 tui tests pass.

Commit: https://github.com/Acture/codex/commit/e087032

Patch (for easy cherry-pick):

@@ tui/src/bottom_pane/approval_overlay.rs @@
-                    "Yes, just this once".to_string()
+                    "Allow just this once".to_string()
-                    "Yes, proceed".to_string()
+                    "Allow and proceed".to_string()
-                        "Yes, and don't ask again for commands that start with `{rendered_prefix}`"
+                        "Allow, and don't ask again for commands that start with `{rendered_prefix}`"
-                    "Yes, and allow this host for this conversation".to_string()
+                    "Allow this host for this conversation".to_string()
-                    "Yes, and allow these permissions for this session".to_string()
+                    "Allow these permissions for this session".to_string()
-                    "Yes, and don't ask again for this command in this session".to_string()
+                    "Allow for this session".to_string()
-                        "No, and block this host in the future".to_string(),
+                        "Deny, and block this host in the future".to_string(),
-                label: "No, continue without running it".to_string(),
+                label: "Deny, continue without running it".to_string(),
-                label: "No, and tell Codex what to do differently".to_string(),
+                label: "Deny, and tell Codex what to do differently".to_string(),
-            label: "Yes, proceed".to_string(),
+            label: "Allow and proceed".to_string(),
-            label: "Yes, and don't ask again for these files".to_string(),
+            label: "Allow, and don't ask again for these files".to_string(),

Same changes applied to tui_app_server/src/bottom_pane/approval_overlay.rs.

Verified locally: cargo test -p codex-tui and cargo test -p codex-tui-app-server pass. just fix -p codex-tui reports no warnings.

etraut-openai contributor · 1 month ago

This feature request hasn't received enough upvotes, so it is being closed automatically. If you still feel this feature is needed, please open a new feature request with an updated title and description.