Digest / November 14, 2025
Codex Issues Digest — November 14, 2025
32 new issues filed ·
37 resolved in openai/codex (UTC day).
New issues (32)
- Streaming HTTP MCP server for Codex
- Version 0.58 displays all our business account usage as full
- ALIVE AI
- Taking a long time / doesn't answer
- failing to apply code
- Fails to see stderr output (of `swift test`)
- Can't acces codex in CLI, and vscode extension
- Honor config.toml symlinks when creating or updating config
- model_reasoning_effort is Not Supported in gpt-5.1-codex
- MCP Playwright & chrome-devtools doesn't work in Codex 5.1, GPT 5.1 Bug
- is headless workaround still valid?
- Codex doesn't see failed state of process
- Annoying "thinking" text effect
- learning mode
- Usage seems to have been reduced or there is another bug around counting usage
- Stream Disconnection
- Codex process errored: Codex process is not available
- Display the shortcut for transcript mode next to long commands
- Support key not only env_key for model_provider
- Codex CLI hangs when configured to use codex mcp-server as an MCP server
- Landlock: fallback to supported ABI or exit with error
- ChatGPT 5.1 Codex does not follow directions in AGENTS.md file.
- Sandbox mode always set to readonly even if set to workspace_write in config.toml
- Not Cool To Be Sneaky and Cut Down Iimits
- Azure provider: config.toml model changed after gpt-5.1-codex release, causing DeploymentNotFound errors
- Codex Extension will not / almost never completes a single prompt on new update
- Shows Plus instead of Pro
- Model is using `rg` wrong when searching for multiple patterns at the same time.
- Can't open link www.openai.com/index/gpt-5-1 from changelog from JetBrains IDE
- Sandbox Meta Data
- Read-only mode fails to run git diff - likely Regression
- Using Skills command line function makes codex very slow on first turn
Resolved issues (37)
- Version 0.58 displays all our business account usage as full
- ALIVE AI
- “Update available” shown before Homebrew cask is updated, causing misleading upgrade message
- Can't acces codex in CLI, and vscode extension
- Feature request: built-in notifications for CLI responses
- code can not revert to the any previous version
- MCP servers in json form
- Need to adapt the mcp Server configuration in Json format
- Timeout when starting a new conversation (newConversation); switching model works
- Bug Report: Excessive Processing Time - 9 Hours
- Streaming HTTP MCP server for Codex
- Feature Request: resume the latest session in the current directory (follow-up to #4342)
- Make history rewind scoped per cwd
- Unable to authenticate Codex extension in Visual Studio Code when WSL is enabled
- Why does the CLI work normally but the IDE plugin shows a 401 error?
- "Allow every time" is asked every time
- Allowed tools settings for non-interactive mode?
- Image in WSL do not work.
- Sticky task summary header + tab title status per agent/terminal
- Are there any features such as adding global codebase indexes?
- Let users rename sessions directly from the resume picker
- Do not indent output
- Support VIM editor commands in CLI text input
- On windows terminal when ssh to a linux host it render error
- cache miss can cause higher usage towards limit than expected
- Codex does not surface 429 Too Many Requests responses from model providers
- How to turn back stats on?
- Squashing commits invokes git rebase -i and hangs
- Codex runs consistently into a stream error when using /compact
- Why does codex keep writing to home directory?
- is headless workaround still valid?
- Stream Disconnection
- When logged in with ChatGPT Plus/Pro while also supplying an API key, automatically fall back to using API credits once ChatGPT credits run out
- Support key not only env_key for model_provider
- Not Cool To Be Sneaky and Cut Down Iimits
- Wrong button text colors on VS Code Light+ theme
- "Don't Trust" option blocks all writes instead of requesting approval