We were assembling the skill roots in two different places, and the
admin root was missing in one of them. This change centralizes root
selection into a helper so both paths stay in sync.
## Description
Introduced `ExternalSandbox` policy to cover use case when sandbox
defined by outside environment, effectively it translates to
`SandboxMode#DangerFullAccess` for file system (since sandbox configured
on container level) and configurable `network_access` (either Restricted
or Enabled by outside environment).
as example you can configure `ExternalSandbox` policy as part of
`sendUserTurn` v1 app_server API:
```
{
"conversationId": <id>,
"cwd": <cwd>,
"approvalPolicy": "never",
"sandboxPolicy": {
"type": ""external-sandbox",
"network_access": "enabled"/"restricted"
},
"model": <model>,
"effort": <effort>,
....
}
```
https://github.com/openai/codex/pull/8235 introduced `ConfigBuilder` and
this PR updates all call non-test call sites to use it instead of
`Config::load_from_base_config_with_overrides()`.
This is important because `load_from_base_config_with_overrides()` uses
an empty `ConfigRequirements`, which is a reasonable default for testing
so the tests are not influenced by the settings on the host. This method
is now guarded by `#[cfg(test)]` so it cannot be used by business logic.
Because `ConfigBuilder::build()` is `async`, many of the test methods
had to be migrated to be `async`, as well. On the bright side, this made
it possible to eliminate a bunch of `block_on_future()` stuff.
Historically, `accept_elicitation_for_prompt_rule()` was flaky because
we were using a notification to update the sandbox followed by a `shell`
tool request that we expected to be subject to the new sandbox config,
but because [rmcp](https://crates.io/crates/rmcp) MCP servers delegate
each incoming message to a new Tokio task, messages are not guaranteed
to be processed in order, so sometimes the `shell` tool call would run
before the notification was processed.
Prior to this PR, we relied on a generous `sleep()` between the
notification and the request to reduce the change of the test flaking
out.
This PR implements a proper fix, which is to use a _request_ instead of
a notification for the sandbox update so that we can wait for the
response to the sandbox request before sending the request to the
`shell` tool call. Previously, `rmcp` did not support custom requests,
but I fixed that in
https://github.com/modelcontextprotocol/rust-sdk/pull/590, which made it
into the `0.12.0` release (see #8288).
This PR updates `shell-tool-mcp` to expect
`"codex/sandbox-state/update"` as a _request_ instead of a notification
and sends the appropriate ack. Note this behavior is tied to our custom
`codex/sandbox-state` capability, which Codex honors as an MCP client,
which is why `core/src/mcp_connection_manager.rs` had to be updated as
part of this PR, as well.
This PR also updates the docs at `shell-tool-mcp/README.md`.
This implements the new config design where config _requirements_ are
loaded separately (and with a special schema) as compared to config
_settings_. In particular, on UNIX, with this PR, you could define
`/etc/codex/requirements.toml` with:
```toml
allowed_approval_policies = ["never", "on-request"]
```
to enforce that `Config.approval_policy` must be one of those two values
when Codex runs.
We plan to expand the set of things that can be restricted by
`/etc/codex/requirements.toml` in short order.
Note that requirements can come from several sources:
- new MDM key on macOS (not implemented yet)
- `/etc/codex/requirements.toml`
- re-interpretation of legacy MDM key on macOS
(`com.openai.codex/config_toml_base64`)
- re-interpretation of legacy `/etc/codex/managed_config.toml`
So our resolution strategy is to load TOML data from those sources, in
order. Later TOMLs are "merged" into previous TOMLs, but any field that
is already set cannot be overwritten. See
`ConfigRequirementsToml::merge_unset_fields()`.
# Terminal Detection Metadata for Per-Terminal Scroll Scaling
## Summary
Expand terminal detection into structured metadata (`TerminalInfo`) with
multiplexer awareness, plus a testable environment shim and
characterization tests.
## Context / Motivation
- TUI2 owns its viewport and scrolling model (see
`codex-rs/tui2/docs/tui_viewport_and_history.md`), so scroll behavior
must be consistent across terminals and independent of terminal
scrollback quirks.
- Prior investigations show mouse wheel scroll deltas vary noticeably by
terminal. To tune scroll scaling (line increments per wheel tick) we
need reliable terminal identification, including when running inside
tmux/zellij.
- tmux is especially tricky because it can mask the underlying terminal;
we now consult `tmux display-message` client termtype/name to attribute
sessions to the actual terminal rather than tmux itself.
- This remains backwards compatible with the existing OpenTelemetry
user-agent token because `user_agent()` is still derived from the same
environment signals (now via `TerminalInfo`).
## Changes
- Introduce `TerminalInfo`, `TerminalName`, and `Multiplexer` with
`TERM_PROGRAM`/`TERM`/multiplexer detection and user-agent formatting in
`codex-rs/core/src/terminal.rs`.
- Add an injectable `Environment` trait + `FakeEnvironment` for testing,
and comprehensive characterization tests covering known terminals, tmux
client termtype/name, and zellij.
- Document module usage and detection order; update `terminal_info()` to
be the primary interface for callers.
## Testing
- `cargo test -p codex-core terminal::tests`
- manually checked ghostty, iTerm2, Terminal.app, vscode, tmux, zellij,
Warp, alacritty, kitty.
```
2025-12-18T07:07:49.191421Z INFO Detected terminal info terminal=TerminalInfo { name: Iterm2, term_program: Some("iTerm.app"), version: Some("3.6.6"), term: None, multiplexer: None }
2025-12-18T07:07:57.991776Z INFO Detected terminal info terminal=TerminalInfo { name: AppleTerminal, term_program: Some("Apple_Terminal"), version: Some("455.1"), term: None, multiplexer: None }
2025-12-18T07:08:07.732095Z INFO Detected terminal info terminal=TerminalInfo { name: WarpTerminal, term_program: Some("WarpTerminal"), version: Some("v0.2025.12.10.08.12.stable_03"), term: None, multiplexer: None }
2025-12-18T07:08:24.860316Z INFO Detected terminal info terminal=TerminalInfo { name: Kitty, term_program: None, version: None, term: None, multiplexer: None }
2025-12-18T07:08:38.302761Z INFO Detected terminal info terminal=TerminalInfo { name: Alacritty, term_program: None, version: None, term: None, multiplexer: None }
2025-12-18T07:08:50.887748Z INFO Detected terminal info terminal=TerminalInfo { name: VsCode, term_program: Some("vscode"), version: Some("1.107.1"), term: None, multiplexer: None }
2025-12-18T07:10:01.309802Z INFO Detected terminal info terminal=TerminalInfo { name: WezTerm, term_program: Some("WezTerm"), version: Some("20240203-110809-5046fc22"), term: None, multiplexer: None }
2025-12-18T08:05:17.009271Z INFO Detected terminal info terminal=TerminalInfo { name: Ghostty, term_program: Some("ghostty"), version: Some("1.2.3"), term: None, multiplexer: None }
2025-12-18T08:05:23.819973Z INFO Detected terminal info terminal=TerminalInfo { name: Ghostty, term_program: Some("ghostty"), version: Some("1.2.3"), term: Some("xterm-ghostty"), multiplexer: Some(Tmux { version: Some("3.6a") }) }
2025-12-18T08:05:35.572853Z INFO Detected terminal info terminal=TerminalInfo { name: Ghostty, term_program: Some("ghostty"), version: Some("1.2.3"), term: None, multiplexer: Some(Zellij) }
```
## Notes / Follow-ups
- Next step is to wire `TerminalInfo` into TUI2’s scroll scaling
configuration and add a per-terminal tuning table.
- The log output in TUI2 helps validate real-world detection before
applying behavior changes.
# External (non-OpenAI) Pull Request Requirements
Before opening this Pull Request, please read the dedicated
"Contributing" markdown file or your PR may be closed:
https://github.com/openai/codex/blob/main/docs/contributing.md
If your PR conforms to our contribution guidelines, replace this text
with a detailed and high quality description of your changes.
Include a link to a bug report or enhancement request.
This is a significant change to how layers of configuration are applied.
In particular, the `ConfigLayerStack` now has two important fields:
- `layers: Vec<ConfigLayerEntry>`
- `requirements: ConfigRequirements`
We merge `TomlValue`s across the layers, but they are subject to
`ConfigRequirements` before creating a `Config`.
How I would review this PR:
- start with `codex-rs/app-server-protocol/src/protocol/v2.rs` and note
the new variants added to the `ConfigLayerSource` enum:
`LegacyManagedConfigTomlFromFile` and `LegacyManagedConfigTomlFromMdm`
- note that `ConfigLayerSource` now has a `precedence()` method and
implements `PartialOrd`
- `codex-rs/core/src/config_loader/layer_io.rs` is responsible for
loading "admin" preferences from `/etc/codex/managed_config.toml` and
MDM. Because `/etc/codex/managed_config.toml` is now deprecated in favor
of `/etc/codex/requirements.toml` and `/etc/codex/config.toml`, we now
include some extra information on the `LoadedConfigLayers` returned in
`layer_io.rs`.
- `codex-rs/core/src/config_loader/mod.rs` has major changes to
`load_config_layers_state()`, which is what produces `ConfigLayerStack`.
The docstring has the new specification and describes the various layers
that will be loaded and the precedence order.
- It uses the information from `LoaderOverrides` "twice," both in the
spirit of legacy support:
- We use one instances to derive an instance of `ConfigRequirements`.
Currently, the only field in `managed_config.toml` that contributes to
`ConfigRequirements` is `approval_policy`. This PR introduces
`Constrained::allow_only()` to support this.
- We use a clone of `LoaderOverrides` to derive
`ConfigLayerSource::LegacyManagedConfigTomlFromFile` and
`ConfigLayerSource::LegacyManagedConfigTomlFromMdm` layers, as
appropriate. As before, this ends up being a "best effort" at enterprise
controls, but is enforcement is not guaranteed like it is for
`ConfigRequirements`.
- Now we only create a "user" layer if `$CODEX_HOME/config.toml` exists.
(Previously, a user layer was always created for `ConfigLayerStack`.)
- Similarly, we only add a "session flags" layer if there are CLI
overrides.
- `config_loader/state.rs` contains the updated implementation for
`ConfigLayerStack`. Note the public API is largely the same as before,
but the implementation is quite different. We leverage the fact that
`ConfigLayerSource` is now `PartialOrd` to ensure layers are in the
correct order.
- A `Config` constructed via `ConfigBuilder.build()` will use
`load_config_layers_state()` to create the `ConfigLayerStack` and use
the associated `ConfigRequirements` when constructing the `Config`
object.
- That said, a `Config` constructed via
`Config::load_from_base_config_with_overrides()` does _not_ yet use
`ConfigBuilder`, so it creates a `ConfigRequirements::default()` instead
of loading a proper `ConfigRequirements`. I will fix this in a
subsequent PR.
Then the following files are mostly test changes:
```
codex-rs/app-server/tests/suite/v2/config_rpc.rs
codex-rs/core/src/config/service.rs
codex-rs/core/src/config_loader/tests.rs
```
Again, because we do not always include "user" and "session flags"
layers when the contents are empty, `ConfigLayerStack` sometimes has
fewer layers than before (and the precedence order changed slightly),
which is the main reason integration tests changed.
# External (non-OpenAI) Pull Request Requirements
Before opening this Pull Request, please read the dedicated
"Contributing" markdown file or your PR may be closed:
https://github.com/openai/codex/blob/main/docs/contributing.md
If your PR conforms to our contribution guidelines, replace this text
with a detailed and high quality description of your changes.
Include a link to a bug report or enhancement request.
Introduce `ConfigBuilder` as an alternative to our existing `Config`
constructors.
I noticed that the existing constructors,
`Config::load_with_cli_overrides()` and
`Config::load_with_cli_overrides_and_harness_overrides()`, did not take
`codex_home` as a parameter, which can be a problem.
Historically, when Codex was purely a CLI, we wanted to be extra sure
that the creation of `codex_home` was always done via
`find_codex_home()`, so we did not expose `codex_home` as a parameter
when creating `Config` in business logic. But in integration tests,
`codex_home` nearly always needs to be configured (as a temp directory),
which is why callers would have to go through
`Config::load_from_base_config_with_overrides()` instead.
Now that the Codex harness also functions as an app server, which could
conceivably load multiple threads where `codex_home` is parameterized
differently in each one, I think it makes sense to make this
configurable. Going to a builder pattern makes it more flexible to
ensure an arbitrary permutation of options can be set when constructing
a `Config` while using the appropriate defaults for the options that
aren't set explicitly.
Ultimately, I think this should make it possible for us to make
`Config::load_from_base_config_with_overrides()` private because all
integration tests should be able to leverage `ConfigBuilder` instead.
Though there could be edge cases, so I'll pursue that migration after we
get through the current config overhaul.
---
[//]: # (BEGIN SAPLING FOOTER)
Stack created with [Sapling](https://sapling-scm.com). Best reviewed
with [ReviewStack](https://reviewstack.dev/openai/codex/pull/8235).
* #8237
* __->__ #8235
1. Remove PUBLIC skills and introduce SYSTEM skills embedded in the
binary and installed into $CODEX_HOME/skills/.system at startup.
2. Skills are now always enabled (feature flag removed).
3. Update skills/list to accept forceReload and plumb it through (not
used by clients yet).
# External (non-OpenAI) Pull Request Requirements
Before opening this Pull Request, please read the dedicated
"Contributing" markdown file or your PR may be closed:
https://github.com/openai/codex/blob/main/docs/contributing.md
If your PR conforms to our contribution guidelines, replace this text
with a detailed and high quality description of your changes.
Include a link to a bug report or enhancement request.
This PR does various types of cleanup before I can proceed with more
ambitious changes to config loading.
First, I noticed duplicated code across these two methods:
774bd9e432/codex-rs/core/src/config/mod.rs (L314-L324)774bd9e432/codex-rs/core/src/config/mod.rs (L334-L344)
This has now been consolidated in
`load_config_as_toml_with_cli_overrides()`.
Further, I noticed that `Config::load_with_cli_overrides()` took two
similar arguments:
774bd9e432/codex-rs/core/src/config/mod.rs (L308-L311)
The difference between `cli_overrides` and `overrides` was not
immediately obvious to me. At first glance, it appears that one should
be able to be expressed in terms of the other, but it turns out that
some fields of `ConfigOverrides` (such as `cwd` and
`codex_linux_sandbox_exe`) are, by design, not configurable via a
`.toml` file or a command-line `--config` flag.
That said, I discovered that many callers of
`Config::load_with_cli_overrides()` were passing
`ConfigOverrides::default()` for `overrides`, so I created two separate
methods:
- `Config::load_with_cli_overrides(cli_overrides: Vec<(String,
TomlValue)>)`
- `Config::load_with_cli_overrides_and_harness_overrides(cli_overrides:
Vec<(String, TomlValue)>, harness_overrides: ConfigOverrides)`
The latter has a long name, as it is _not_ what should be used in the
common case, so the extra typing is designed to draw attention to this
fact. I tried to update the existing callsites to use the shorter name,
where possible.
Further, in the cases where `ConfigOverrides` is used, usually only a
limited subset of fields are actually set, so I updated the declarations
to leverage `..Default::default()` where possible.
# External (non-OpenAI) Pull Request Requirements
Before opening this Pull Request, please read the dedicated
"Contributing" markdown file or your PR may be closed:
https://github.com/openai/codex/blob/main/docs/contributing.md
If your PR conforms to our contribution guidelines, replace this text
with a detailed and high quality description of your changes.
Include a link to a bug report or enhancement request.
We should not have any `PathBuf` fields in `ConfigToml` or any of the
transitive structs we include, as we should use `AbsolutePathBuf`
instead so that we do not have to keep track of the file from which
`ConfigToml` was loaded such that we need it to resolve relative paths
later when the values of `ConfigToml` are used.
I only found two instances of this: `experimental_instructions_file` and
`experimental_compact_prompt_file`. Incidentally, when these were
specified as relative paths, they were resolved against `cwd` rather
than `config.toml`'s parent, which seems wrong to me. I changed the
behavior so they are resolved against the parent folder of the
`config.toml` being parsed, which we get "for free" due to the
introduction of `AbsolutePathBufGuard ` in
https://github.com/openai/codex/pull/7796.
While it is not great to change the behavior of a released feature,
these fields are prefixed with `experimental_`, which I interpret to
mean we have the liberty to change the contract.
For reference:
- `experimental_instructions_file` was introduced in
https://github.com/openai/codex/pull/1803
- `experimental_compact_prompt_file` was introduced in
https://github.com/openai/codex/pull/5959
The problem is that the `tokio` task own an `Arc` reference of the
session and that this task only exit with the broadcast channel get
closed. But this never get closed if the session is not dropped. So it's
a snake biting his tail basically
The most notable result was that non of the `Drop` implementation were
triggered (temporary files, shell snapshots, session cleaning etc etc)
when closing the session (through a `/new` for example)
The fix is just to weaken the `Arc` and upgrade it on the fly
Constrain `approval_policy` through new `admin_policy` config.
This PR will:
1. Add a `admin_policy` section to config, with a single field (for now)
`allowed_approval_policies`. This list constrains the set of
user-settable `approval_policy`s.
2. Introduce a new `Constrained<T>` type, which combines a current value
and a validator function. The validator function ensures disallowed
values are not set.
3. Change the type of `approval_policy` on `Config` and
`SessionConfiguration` from `AskForApproval` to
`Constrained<AskForApproval>`. The validator function is set by the
values passed into `allowed_approval_policies`.
4. `GenericDisplayRow`: add a `disabled_reason: Option<String>`. When
set, it disables selection of the value and indicates as such in the
menu. This also makes it unselectable with arrow keys or numbers. This
is used in the `/approvals` menu.
Follow ups are:
1. Do the same thing to `sandbox_policy`.
2. Propagate the allowed set of values through app-server for the
extension (though already this should prevent app-server from setting
this values, it's just that we want to disable UI elements that are
unsettable).
Happy to split this PR up if you prefer, into the logical numbered areas
above. Especially if there are parts we want to gavel on separately
(e.g. admin_policy).
Disabled full access:
<img width="1680" height="380" alt="image"
src="https://github.com/user-attachments/assets/1fb61c8c-1fcb-4dc4-8355-2293edb52ba0"
/>
Disabled `--yolo` on startup:
<img width="749" height="76" alt="image"
src="https://github.com/user-attachments/assets/0a1211a0-6eb1-40d6-a1d7-439c41e94ddb"
/>
CODEX-4087
This attempts to tighten up the types related to "config layers."
Currently, `ConfigLayerEntry` is defined as follows:
bef36f4ae7/codex-rs/core/src/config_loader/state.rs (L19-L25)
but the `source` field is a bit of a lie, as:
- for `ConfigLayerName::Mdm`, it is
`"com.openai.codex/config_toml_base64"`
- for `ConfigLayerName::SessionFlags`, it is `"--config"`
- for `ConfigLayerName::User`, it is `"config.toml"` (just the file
name, not the path to the `config.toml` on disk that was read)
- for `ConfigLayerName::System`, it seems like it is usually
`/etc/codex/managed_config.toml` in practice, though on Windows, it is
`%CODEX_HOME%/managed_config.toml`:
bef36f4ae7/codex-rs/core/src/config_loader/layer_io.rs (L84-L101)
All that is to say, in three out of the four `ConfigLayerName`, `source`
is a `PathBuf` that is not an absolute path (or even a true path).
This PR tries to uplevel things by eliminating `source` from
`ConfigLayerEntry` and turning `ConfigLayerName` into a disjoint union
named `ConfigLayerSource` that has the appropriate metadata for each
variant, favoring the use of `AbsolutePathBuf` where appropriate:
```rust
pub enum ConfigLayerSource {
/// Managed preferences layer delivered by MDM (macOS only).
#[serde(rename_all = "camelCase")]
#[ts(rename_all = "camelCase")]
Mdm { domain: String, key: String },
/// Managed config layer from a file (usually `managed_config.toml`).
#[serde(rename_all = "camelCase")]
#[ts(rename_all = "camelCase")]
System { file: AbsolutePathBuf },
/// Session-layer overrides supplied via `-c`/`--config`.
SessionFlags,
/// User config layer from a file (usually `config.toml`).
#[serde(rename_all = "camelCase")]
#[ts(rename_all = "camelCase")]
User { file: AbsolutePathBuf },
}
```
1. Adds SkillScope::Public end-to-end (core + protocol) and loads skills
from the public cache directory
2. Improves repo skill discovery by searching upward for the nearest
.codex/skills within a git repo
3. Deduplicates skills by name with deterministic ordering to avoid
duplicates across sources
4. Fixes garbled “Skill errors” overlay rendering by preventing pending
history lines from being injected during the modal
5. Updates the project docs “Skills” intro wording to avoid hardcoded
paths
This reverts commit 291b54a762.
This commit was intended to prevent the model from making code changes
during `/review`, which is sometimes does. Unfortunately, it has other
unintended side effects that cause `/review` to fail in a variety of
ways. See #8115 and #7815. We've therefore decided to revert this
change.
