## Overview
Adds LM Studio OSS support. Closes#1883
### Changes
This PR enhances the behavior of `--oss` flag to support LM Studio as a
provider. Additionally, it introduces a new flag`--local-provider` which
can take in `lmstudio` or `ollama` as values if the user wants to
explicitly choose which one to use.
If no provider is specified `codex --oss` will auto-select the provider
based on whichever is running.
#### Additional enhancements
The default can be set using `oss-provider` in config like:
```
oss_provider = "lmstudio"
```
For non-interactive users, they will need to either provide the provider
as an arg or have it in their `config.toml`
### Notes
For best performance, [set the default context
length](https://lmstudio.ai/docs/app/advanced/per-model) for gpt-oss to
the maximum your machine can support
---------
Co-authored-by: Matt Clayton <matt@lmstudio.ai>
Co-authored-by: Eric Traut <etraut@openai.com>
## Summary
Fixes streaming issue where Claude models return only 1-4 characters
instead of full responses when used through certain API
providers/proxies.
## Environment
- **OS**: Windows
- **Models affected**: Claude models (e.g., claude-haiku-4-5-20251001)
- **API Provider**: AAAI API proxy (https://api.aaai.vip/v1)
- **Working models**: GLM, Google models work correctly
## Problem
When using Claude models in both TUI and exec modes, only 1-4 characters
are displayed despite the backend receiving the full response. Debug
logs revealed that some API providers send SSE chunks with an empty
string finish_reason during active streaming, rather than null or
omitting the field entirely.
The current code treats any non-null finish_reason as a termination
signal, causing the stream to exit prematurely after the first chunk.
The problematic chunks contain finish_reason with an empty string
instead of null.
## Solution
Fix empty finish_reason handling in chat_completions.rs by adding a
check to only process non-empty finish_reason values. This ensures empty
strings are ignored and streaming continues normally.
## Testing
- Tested on Windows with Claude Haiku model via AAAI API proxy
- Full responses now received and displayed correctly in both TUI and
exec modes
- Other models (GLM, Google) continue to work as expected
- No regression in existing functionality
## Impact
- Improves compatibility with API providers that send empty
finish_reason during streaming
- Enables Claude models to work correctly in Windows environment
- No breaking changes to existing functionality
## Related Issues
This fix resolves the issue where Claude models appeared to return
incomplete responses. The root cause was identified as a compatibility
issue in parsing SSE responses from certain API providers/proxies,
rather than a model-specific problem. This change improves overall
robustness when working with various API endpoints.
---------
Co-authored-by: Eric Traut <etraut@openai.com>
This PR does the following:
- Add compact prefix to the summary
- Change the compaction prompt
- Allow multiple compaction for long running tasks
- Filter out summary messages on the following compaction
Considerations:
- Filtering out the summary message isn't the most clean
- Theoretically, we can end up in infinite compaction loop if the user
messages > compaction limit . However, that's not possible in today's
code because we have hard cap on user messages.
- We need to address having multiple user messages because it confuses
the model.
Testing:
- Making sure that after compact we always end up with one user message
(task) and one summary, even on multiple compaction.
Fixes#4940Fixes#4892
When selecting "No, ask me to approve edits and commands" during
onboarding, the code wasn't applying the correct approval policy,
causing Codex to block all write operations instead of requesting
approval.
This PR fixes the issue by persisting the "DontTrust" decision in
config.toml as `trust_level = "untrusted"` and handling it in the
sandbox and approval policy logic, so Codex correctly asks for approval
before making changes.
## Before (bug)
<img width="709" height="500" alt="bef"
src="https://github.com/user-attachments/assets/5aced26d-d810-4754-879a-89d9e4e0073b"
/>
## After (fixed)
<img width="713" height="359" alt="aft"
src="https://github.com/user-attachments/assets/9887bbcb-a9a5-4e54-8e76-9125a782226b"
/>
---------
Co-authored-by: Eric Traut <etraut@openai.com>
For better caching performance all output items should be rendered in
the order they were produced before all new input items (for example,
all function_call before all function_call_output).
## Summary
- default the `tui.notifications` setting to enabled so desktop
notifications work out of the box
- update configuration tests and documentation to reflect the new
default
## Testing
- `cargo test -p codex-core` *(fails:
`exec::tests::kill_child_process_group_kills_grandchildren_on_timeout`
is flaky in this sandbox because the spawned grandchild process stays
alive)*
- `cargo test -p codex-core
exec::tests::kill_child_process_group_kills_grandchildren_on_timeout`
*(fails: same sandbox limitation as above)*
------
[Codex
Task](https://chatgpt.com/codex/tasks/task_i_69166f811144832c9e8aaf8ee2642373)
core event to app server event mapping:
1. `codex/event/reasoning_content_delta` ->
`item/reasoning/summaryTextDelta`.
2. `codex/event/reasoning_raw_content_delta` ->
`item/reasoning/textDelta`
3. `codex/event/agent_message_content_delta` →
`item/agentMessage/delta`.
4. `codex/event/agent_reasoning_section_break` ->
`item/reasoning/summaryPartAdded`.
Also added a change in core to pass down content index, summary index
and item id from events.
Tested with the `git checkout owen/app_server_test_client && cargo run
-p codex-app-server-test-client -- send-message-v2 "hello"` and verified
that new events are emitted correctly.
We've received many reports of codex hanging when calling certain tools.
[Here](https://github.com/openai/codex/issues/3204) is one example. This
is likely a major cause. The problem occurs when
`consume_truncated_output` waits for `stdout` and `stderr` to be closed
once the child process terminates. This normally works fine, but it
doesn't handle the case where the child has spawned grandchild processes
that inherits `stdout` and `stderr`.
The fix was originally written by @md-oai in [this
PR](https://github.com/openai/codex/pull/1852), which has gone stale.
I've copied the original fix (which looks sound to me) and added an
integration test to prevent future regressions.
- Introducing a screen to inform users of model changes.
- Config name is being passed to be able to reuse this component in the
future for future models
This PR addresses https://github.com/openai/codex/issues/6360. The root
problem is that the TUI was directly loading the `auth.json` file to
access the auth information. It should instead be using the AuthManager,
which records the current auth information. The `auth.json` file can be
overwritten at any time by other instances of the CLI or extension, so
its information can be out of sync with the current instance. The
`/status` command should always report the auth information associated
with the current instance.
An alternative fix for this bug was submitted by @chojs23 in [this
PR](https://github.com/openai/codex/pull/6495). That approach was only a
partial fix.
This adds support for a new variant of the shell tool behind a flag. To
test, run `codex` with `--enable shell_command_tool`, which will
register the tool with Codex under the name `shell_command` that accepts
the following shape:
```python
{
command: str
workdir: str | None,
timeout_ms: int | None,
with_escalated_permissions: bool | None,
justification: str | None,
}
```
This is comparable to the existing tool registered under
`shell`/`container.exec`. The primary difference is that it accepts
`command` as a `str` instead of a `str[]`. The `shell_command` tool
executes by running `execvp(["bash", "-lc", command])`, though the exact
arguments to `execvp(3)` depend on the user's default shell.
The hypothesis is that this will simplify things for the model. For
example, on Windows, instead of generating:
```json
{"command": ["pwsh.exe", "-NoLogo", "-Command", "ls -Name"]}
```
The model could simply generate:
```json
{"command": "ls -Name"}
```
As part of this change, I extracted some logic out of `user_shell.rs` as
`Shell::derive_exec_args()` so that it can be reused in
`codex-rs/core/src/tools/handlers/shell.rs`. Note the original code
generated exec arg lists like:
```javascript
["bash", "-lc", command]
["zsh", "-lc", command]
["pwsh.exe", "-NoProfile", "-Command", command]
```
Using `-l` for Bash and Zsh, but then specifying `-NoProfile` for
PowerShell seemed inconsistent to me, so I changed this in the new
implementation while also adding a `use_login_shell: bool` option to
make this explicit. If we decide to add a `login: bool` to
`ShellCommandToolCallParams` like we have for unified exec:
https://github.com/openai/codex/blob/807e2c27f0a9f2e85c50e7e6df5533f0d9b853c7/codex-rs/core/src/tools/handlers/unified_exec.rs#L33-L34
Then this should make it straightforward to support.
Unified exec isn't working on Linux because we don't provide the correct
arg0.
The library we use for pty management doesn't allow setting arg0
separately from executable. Use the same aliasing strategy we use for
`apply_patch` for `codex-linux-sandbox`.
Use `#[ctor]` hack to dispatch codex-linux-sandbox calls.
Addresses https://github.com/openai/codex/issues/6450
## Summary
- add a `hide_rate_limit_model_nudge` notice flag plus config edit
plumbing so the rate limit reminder preference is persisted and
documented
- extend the chat widget prompt with a "never show again" option, and
wire new app events so selecting it hides future nudges immediately and
writes the config
- add unit coverage and refresh the snapshot for the three-option prompt
## Testing
- `just fmt`
- `just fix -p codex-tui`
- `just fix -p codex-core`
- `cargo test -p codex-tui`
- `cargo test -p codex-core` *(fails at
`exec::tests::kill_child_process_group_kills_grandchildren_on_timeout`:
grandchild process still alive)*
------
[Codex
Task](https://chatgpt.com/codex/tasks/task_i_6910d7f407748321b2661fc355416994)
## Summary
- launch shell tool processes in their own process group so Codex owns
the full tree
- on timeout or ctrl-c, send SIGKILL to the process group before
terminating the tracked child
- document that the default shell/unified_exec timeout remains 1000 ms
## Original Bug
Long-lived shell tool commands hang indefinitely because the timeout
handler only terminated the direct child process; any grandchildren it
spawned kept running and held the PTY open, preventing Codex from
regaining control.
## Repro Original Bug
Install next.js and run `next dev` (which is a long-running shell
process with children). On openai:main, it will cause the agent to
permanently get stuck here until human intervention. On this branch,
this command will be terminated successfully after timeout_ms which will
unblock the agent. This is a critical fix for unmonitored / lightly
monitored agents that don't have immediate human observation to unblock
them.
---------
Co-authored-by: Michael Bolin <mbolin@openai.com>
Co-authored-by: Michael Bolin <bolinfest@gmail.com>
### Motivation
When Codex is launched from a region where Cloudflare blocks access (for
example, Russia), the CLI currently dumps Cloudflare’s entire HTML error
page. This isn’t actionable and makes it hard for users to understand
what happened. We want to detect the Cloudflare block and show a
concise, user-friendly explanation instead.
### What Changed
- Added CLOUDFLARE_BLOCKED_MESSAGE and a friendly_message() helper to
UnexpectedResponseError. Whenever we see a 403 whose body contains the
Cloudflare block notice, we now emit a single-line message (Access
blocked by Cloudflare…) while preserving the HTTP status and request id.
All other responses keep the original behaviour.
- Added two focused unit tests:
- unexpected_status_cloudflare_html_is_simplified ensures the Cloudflare
HTML case yields the friendly message.
- unexpected_status_non_html_is_unchanged confirms plain-text 403s still
return the raw body.
### Testing
- cargo build -p codex-cli
- cargo test -p codex-core
- just fix -p codex-core
- cargo test --all-features
---------
Co-authored-by: Eric Traut <etraut@openai.com>
This PR does two things:
1. add a new function in core that maps the core-internal plan type to
the external plan type;
2. implement account/read that get account status (v2 of
`getAuthStatus`).
This PR makes an "insufficient quota" error fatal so we don't attempt to
retry it multiple times in the agent loop.
We have multiple bug reports from users about intermittent retry
behaviors, and this could explain some of them. With this change, we'll
eliminate the retries and surface a clear error message.
The PR is a nearly identical copy of [this
PR](https://github.com/openai/codex/pull/4837) contributed by
@abimaelmartell. The original PR has gone stale. Rather than wait for
the contributor to resolve merge conflicts, I wanted to get this change
in.
Dylan HurdandGitHub
Ahmed IbrahimandGitHub
jif-oaiandGitHub
Celia ChenandGitHub
Michael BolinandGitHub
zhao-oaiandGitHub
Oliver MannionandGitHub
