Digest / January 16, 2026
Codex Issues Digest — January 16, 2026
27 new issues filed ·
13 resolved in openai/codex (UTC day).
New issues (27)
- Codex CLI becomes unresponsive
- Codex Windows sandbox warning is transient, non-actionable, and causes silent Git failures
- Cannot start Codex CLI since v0.84.0
- Add file undo option to backtrack mode (Ctrl+Enter)
- The Codex CLI has a serious memory leak issue that causes computer crashes.
- Support NO_PROXY / http.noProxy settings in VSCode Extension
- Better distinguish between sessions in `codex resume --all` or `codex resume`
- macOS: LC_ALL forced to C.UTF-8 in codex CLI shell commands (unified exec)
- Compact failed
- Codex suggests incorrect upgrade command: `npm install -g codex` instead of `npm install -g @openai/codex`
- Fix invalid input[].id values containing colons during /review
- WSL: shortcut overlay snapshot tests are non-deterministic
- Feature Request: Add /clear command to clear screen
- Codex cli doesn't see symlinked SKILL.md file in ~/.codex/skills/$skillName/
- Allow us to edit the permission prompt
- Mouse Properties dialog opens unexpectedly on Windows
- Code Quality and Developer Experience Issues
- encoding errado nas acentuações e caracteres especiais para o idioma Português do Brasil
- [FEATURE] Support Claude Code slash command format for compatibility
- fake news usage limit warning
- Expose thinking level param for Codex CLI
- Input broken in TUI when shell_snapshot is enabled (macOS Terminal.app)
- python_multiprocessing_lock_works_under_sandbox fails on Python 3.14; Python 3.12 passes
- Developer role is Incompatible Locally Deployed Model Interface
- Ollama uses CPU after 10k token input to gpt-oss:120b despite model loaded entirely on GPU. Only switches back to GPU after input is less than 10k tokens.
- Unclear interaction between approval mode and file edits
- Stale incremental cache causes 'no field' compilation errors
Resolved issues (13)
- Unable to paste plain text, file path, or anything in Codex CLI V0.81.0
- Sometimes see "Worked for" when job at planning stage
- MCP server never emits TurnComplete event - only threadId returned to caller
- estimate_token_count units bug
- Quota exceeded with 93% available
- Cannot start Codex CLI since v0.84.0
- Codex MCP: Unable to connect to Context7 (Docker startup timeout / npx server fails)
- Codex suggests incorrect upgrade command: `npm install -g codex` instead of `npm install -g @openai/codex`
- encoding errado nas acentuações e caracteres especiais para o idioma Português do Brasil
- Codex CLI slow startup / severe input lag (Possibly since 0.84, possibly related to IO / FileProvider churn)
- Feature Request: Add /clear command to clear screen
- Ollama uses CPU after 10k token input to gpt-oss:120b despite model loaded entirely on GPU. Only switches back to GPU after input is less than 10k tokens.
- Codex cli doesn't see symlinked SKILL.md file in ~/.codex/skills/$skillName/