## Why
`PermissionProfile` should describe filesystem roots as absolute paths
at the type level. Using `PathBuf` in `FileSystemPermissions` made the
shared type too permissive and blurred together three different
deserialization cases:
- skill metadata in `agents/openai.yaml`, where relative paths should
resolve against the skill directory
- app-server API payloads, where callers should have to send absolute
paths
- local tool-call payloads for commands like `shell_command` and
`exec_command`, where `additional_permissions.file_system` may
legitimately be relative to the command `workdir`
This change tightens the shared model without regressing the existing
local command flow.
## What Changed
- changed `protocol::models::FileSystemPermissions` and the app-server
`AdditionalFileSystemPermissions` mirror to use `AbsolutePathBuf`
- wrapped skill metadata deserialization in `AbsolutePathBufGuard`, so
relative permission roots in `agents/openai.yaml` resolve against the
containing skill directory
- kept app-server/API deserialization strict, so relative
`additionalPermissions.fileSystem.*` paths are rejected at the boundary
- restored cwd/workdir-relative deserialization for local tool-call
payloads by parsing `shell`, `shell_command`, and `exec_command`
arguments under an `AbsolutePathBufGuard` rooted at the resolved command
working directory
- simplified runtime additional-permission normalization so it only
canonicalizes and deduplicates absolute roots instead of trying to
recover relative ones later
- updated the app-server schema fixtures, `app-server/README.md`, and
the affected transport/TUI tests to match the final behavior
## Why
Before this change, an escalation approval could say that a command
should be rerun, but it could not carry the sandbox configuration that
should still apply when the escalated command is actually spawned.
That left an unsafe gap in the `zsh-fork` skill path: skill scripts
under `scripts/` that did not declare permissions could be escalated
without a sandbox, and scripts that did declare permissions could lose
their bounded sandbox on rerun or cached session approval.
This PR extends the escalation protocol so approvals can optionally
carry sandbox configuration all the way through execution. That lets the
shell runtime preserve the intended sandbox instead of silently widening
access.
We likely want a single permissions type for this codepath eventually,
probably centered on `Permissions`. For now, the protocol needs to
represent both the existing `PermissionProfile` form and the fuller
`Permissions` form, so this introduces a temporary disjoint union,
`EscalationPermissions`, to carry either one.
Further, this means that today, a skill either:
- does not declare any permissions, in which case it is run using the
default sandbox for the turn
- specifies permissions, in which case the skill is run using that exact
sandbox, which might be more restrictive than the default sandbox for
the turn
We will likely change the skill's permissions to be additive to the
existing permissions for the turn.
## What Changed
- Added `EscalationPermissions` to `codex-protocol` so escalation
requests can carry either a `PermissionProfile` or a full `Permissions`
payload.
- Added an explicit `EscalationExecution` mode to the shell escalation
protocol so reruns distinguish between `Unsandboxed`, `TurnDefault`, and
`Permissions(...)` instead of overloading `None`.
- Updated `zsh-fork` shell reruns to resolve `TurnDefault` at execution
time, which keeps ordinary `UseDefault` commands on the turn sandbox and
preserves turn-level macOS seatbelt profile extensions.
- Updated the `zsh-fork` skill path so a skill with no declared
permissions inherits the conversation's effective sandbox instead of
escalating unsandboxed.
- Updated the `zsh-fork` skill path so a skill with declared permissions
reruns with exactly those permissions, including when a cached session
approval is reused.
## Testing
- Added unit coverage in
`core/src/tools/runtimes/shell/unix_escalation.rs` for the explicit
`UseDefault` / `RequireEscalated` / `WithAdditionalPermissions`
execution mapping.
- Added unit coverage in
`core/src/tools/runtimes/shell/unix_escalation.rs` for macOS seatbelt
extension preservation in both the `TurnDefault` and
explicit-permissions rerun paths.
- Added integration coverage in `core/tests/suite/skill_approval.rs` for
permissionless skills inheriting the turn sandbox and explicit skill
permissions remaining bounded across cached approval reuse.
This reverts commit daf0f03ac8.
# 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.
Summary
- detect skill-invoking shell commands based on the original command
string, request approvals when needed, and cache positive decisions per
session
- keep implicit skill invocation emitted after approval and keep skill
approval decline messaging centralized to the shell handler
- expand and adjust skill approval tests to cover shell-based skill
scripts while matching the new detection expectations
Testing
- Not run (not requested)
## Summary
Introduces the initial implementation of Feature::RequestPermissions.
RequestPermissions allows the model to request that a command be run
inside the sandbox, with additional permissions, like writing to a
specific folder. Eventually this will include other rules as well, and
the ability to persist these permissions, but this PR is already quite
large - let's get the core flow working and go from there!
<img width="1279" height="541" alt="Screenshot 2026-02-15 at 2 26 22 PM"
src="https://github.com/user-attachments/assets/0ee3ec0f-02ec-4509-91a2-809ac80be368"
/>
## Testing
- [x] Added tests
- [x] Tested locally
- [x] Feature
## Summary
Simplify network approvals by removing per-attempt proxy correlation and
moving to session-level approval dedupe keyed by (host, protocol, port).
Instead of encoding attempt IDs into proxy credentials/URLs, we now
treat approvals as a destination policy decision.
- Concurrent calls to the same destination share one approval prompt.
- Different destinations (or same host on different ports) get separate
prompts.
- Allow once approves the current queued request group only.
- Allow for session caches that (host, protocol, port) and auto-allows
future matching requests.
- Never policy continues to deny without prompting.
Example:
- 3 calls:
- a.com (line 443)
- b.com (line 443)
- a.com (line 443)
=> 2 prompts total (a, b), second a waits on the first decision.
- a.com:80 is treated separately from a.com line 443
## Testing
- `just fmt` (in `codex-rs`)
- `cargo test -p codex-core tools::network_approval::tests`
- `cargo test -p codex-core` (unit tests pass; existing
integration-suite failures remain in this environment)
zsh fork PR stack:
- https://github.com/openai/codex/pull/12051
- https://github.com/openai/codex/pull/12052👈
### Summary
This PR introduces a feature-gated native shell runtime path that routes
shell execution through a patched zsh exec bridge, removing MCP-specific
behavior from the shell hot path while preserving existing
CommandExecution lifecycle semantics.
When shell_zsh_fork is enabled, shell commands run via patched zsh with
per-`execve` interception through EXEC_WRAPPER. Core receives wrapper
IPC requests over a Unix socket, applies existing approval policy, and
returns allow/deny before the subcommand executes.
### What’s included
**1) New zsh exec bridge runtime in core**
- Wrapper-mode entrypoint (maybe_run_zsh_exec_wrapper_mode) for
EXEC_WRAPPER invocations.
- Per-execution Unix-socket IPC handling for wrapper requests/responses.
- Approval callback integration using existing core approval
orchestration.
- Streaming stdout/stderr deltas to existing command output event
pipeline.
- Error handling for malformed IPC, denial/abort, and execution
failures.
**2) Session lifecycle integration**
SessionServices now owns a `ZshExecBridge`.
Session startup initializes bridge state; shutdown tears it down
cleanly.
**3) Shell runtime routing (feature-gated)**
When `shell_zsh_fork` is enabled:
- Build execution env/spec as usual.
- Add wrapper socket env wiring.
- Execute via `zsh_exec_bridge.execute_shell_request(...)` instead of
the regular shell path.
- Non-zsh-fork behavior remains unchanged.
**4) Config + feature wiring**
- Added `Feature::ShellZshFork` (under development).
- Added config support for `zsh_path` (optional absolute path to patched
zsh):
- `Config`, `ConfigToml`, `ConfigProfile`, overrides, and schema.
- Session startup validates that `zsh_path` exists/usable when zsh-fork
is enabled.
- Added startup test for missing `zsh_path` failure mode.
**5) Seatbelt/sandbox updates for wrapper IPC**
- Extended seatbelt policy generation to optionally allow outbound
connection to explicitly permitted Unix sockets.
- Wired sandboxing path to pass wrapper socket path through to seatbelt
policy generation.
- Added/updated seatbelt tests for explicit socket allow rule and
argument emission.
**6) Runtime entrypoint hooks**
- This allows the same binary to act as the zsh wrapper subprocess when
invoked via `EXEC_WRAPPER`.
**7) Tool selection behavior**
- ToolsConfig now prefers ShellCommand type when shell_zsh_fork is
enabled.
- Added test coverage for precedence with unified-exec enabled.
### Description
#### Summary
Introduces the core plumbing required for structured network approvals
#### What changed
- Added structured network policy decision modeling in core.
- Added approval payload/context types needed for network approval
semantics.
- Wired shell/unified-exec runtime plumbing to consume structured
decisions.
- Updated related core error/event surfaces for structured handling.
- Updated protocol plumbing used by core approval flow.
- Included small CLI debug sandbox compatibility updates needed by this
layer.
#### Why
establishes the minimal backend foundation for network approvals without
yet changing high-level orchestration or TUI behavior.
#### Notes
- Behavior remains constrained by existing requirements/config gating.
- Follow-up PRs in the stack handle orchestration, UX, and app-server
integration.
---------
Co-authored-by: Codex <199175422+chatgpt-codex-connector[bot]@users.noreply.github.com>
`SandboxPolicy::ReadOnly` previously implied broad read access and could
not express a narrower read surface.
This change introduces an explicit read-access model so we can support
user-configurable read restrictions in follow-up work, while preserving
current behavior today.
It also ensures unsupported backends fail closed for restricted-read
policies instead of silently granting broader access than intended.
## What
- Added `ReadOnlyAccess` in protocol with:
- `Restricted { include_platform_defaults, readable_roots }`
- `FullAccess`
- Updated `SandboxPolicy` to carry read-access configuration:
- `ReadOnly { access: ReadOnlyAccess }`
- `WorkspaceWrite { ..., read_only_access: ReadOnlyAccess }`
- Preserved existing behavior by defaulting current construction paths
to `ReadOnlyAccess::FullAccess`.
- Threaded the new fields through sandbox policy consumers and call
sites across `core`, `tui`, `linux-sandbox`, `windows-sandbox`, and
related tests.
- Updated Seatbelt policy generation to honor restricted read roots by
emitting scoped read rules when full read access is not granted.
- Added fail-closed behavior on Linux and Windows backends when
restricted read access is requested but not yet implemented there
(`UnsupportedOperation`).
- Regenerated app-server protocol schema and TypeScript artifacts,
including `ReadOnlyAccess`.
## Compatibility / rollout
- Runtime behavior remains unchanged by default (`FullAccess`).
- API/schema changes are in place so future config wiring can enable
restricted read access without another policy-shape migration.
This PR adds the following field to `Config`:
```rust
pub network: Option<NetworkProxy>,
```
Though for the moment, it will always be initialized as `None` (this
will be addressed in a subsequent PR).
This PR does the work to thread `network` through to `execute_exec_env()`, `process_exec_tool_call()`, and `UnifiedExecRuntime.run()` to ensure it is available whenever we span a process.
## Summary
This PR introduces a gated Bubblewrap (bwrap) Linux sandbox path. The
curent Linux sandbox path relies on in-process restrictions (including
Landlock). Bubblewrap gives us a more uniform filesystem isolation
model, especially explicit writable roots with the option to make some
directories read-only and granular network controls.
This is behind a feature flag so we can validate behavior safely before
making it the default.
- Added temporary rollout flag:
- `features.use_linux_sandbox_bwrap`
- Preserved existing default path when the flag is off.
- In Bubblewrap mode:
- Added internal retry without /proc when /proc mount is not permitted
by the host/container.
## 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>,
....
}
```
helpful in the future if we want more granularity for requesting
escalated permissions:
e.g when running in readonly sandbox, model can request to escalate to a
sandbox that allows writes
No integration test for now because it would make them flaky. Tracking
it in my todos to add some once we have a clock based system for
integration tests
`process_exec_tool_call()` was taking `SandboxType` as a param, but in
practice, the only place it was constructed was in
`codex_message_processor.rs` where it was derived from the other
`sandbox_policy` param, so this PR inlines the logic that decides the
`SandboxType` into `process_exec_tool_call()`.
---
[//]: # (BEGIN SAPLING FOOTER)
Stack created with [Sapling](https://sapling-scm.com). Best reviewed
with [ReviewStack](https://reviewstack.dev/openai/codex/pull/7122).
* #7112
* __->__ #7122
Previously, we were running into an issue where we would run the `shell`
tool call with a timeout of 10s, but it fired an elicitation asking for
user approval, the time the user took to respond to the elicitation was
counted agains the 10s timeout, so the `shell` tool call would fail with
a timeout error unless the user is very fast!
This PR addresses this issue by introducing a "stopwatch" abstraction
that is used to manage the timeout. The idea is:
- `Stopwatch::new()` is called with the _real_ timeout of the `shell`
tool call.
- `process_exec_tool_call()` is called with the `Cancellation` variant
of `ExecExpiration` because it should not manage its own timeout in this
case
- the `Stopwatch` expiration is wired up to the `cancel_rx` passed to
`process_exec_tool_call()`
- when an elicitation for the `shell` tool call is received, the
`Stopwatch` pauses
- because it is possible for multiple elicitations to arrive
concurrently, it keeps track of the number of "active pauses" and does
not resume until that counter goes down to zero
I verified that I can test the MCP server using
`@modelcontextprotocol/inspector` and specify `git status` as the
`command` with a timeout of 500ms and that the elicitation pops up and I
have all the time in the world to respond whereas previous to this PR,
that would not have been possible.
---
[//]: # (BEGIN SAPLING FOOTER)
Stack created with [Sapling](https://sapling-scm.com). Best reviewed
with [ReviewStack](https://reviewstack.dev/openai/codex/pull/6973).
* #7005
* __->__ #6973
* #6972
This updates `ExecParams` so that instead of taking `timeout_ms:
Option<u64>`, it now takes a more general cancellation mechanism,
`ExecExpiration`, which is an enum that includes a
`Cancellation(tokio_util::sync::CancellationToken)` variant.
If the cancellation token is fired, then `process_exec_tool_call()`
returns in the same way as if a timeout was exceeded.
This is necessary so that in #6973, we can manage the timeout logic
external to the `process_exec_tool_call()` because we want to "suspend"
the timeout when an elicitation from a human user is pending.
---
[//]: # (BEGIN SAPLING FOOTER)
Stack created with [Sapling](https://sapling-scm.com). Best reviewed
with [ReviewStack](https://reviewstack.dev/openai/codex/pull/6972).
* #7005
* #6973
* __->__ #6972
- Use /bin/sh instead of /bin/bash on FreeBSD/OpenBSD in the process
group timeout test to avoid command-not-found failures.
- Accept /usr/local/bin/bash as a valid SHELL path to match common
FreeBSD installations.
- Switch the shell serialization duration test to /bin/sh for improved
portability across Unix platforms.
With this change, `cargo test -p codex-core --lib` runs and passes on
FreeBSD.
## 🐛 Problem
Users running commands with non-ASCII characters (like Russian text
"пример") in Windows/WSL environments experience garbled text in
VSCode's shell preview window, with Unicode replacement characters (�)
appearing instead of the actual text.
**Issue**: https://github.com/openai/codex/issues/6178
## 🔧 Root Cause
The issue was in `StreamOutput<Vec<u8>>::from_utf8_lossy()` method in
`codex-rs/core/src/exec.rs`, which used `String::from_utf8_lossy()` to
convert shell output bytes to strings. This function immediately
replaces any invalid UTF-8 byte sequences with replacement characters,
without attempting to decode using other common encodings.
In Windows/WSL environments, shell output often uses encodings like:
- Windows-1252 (common Windows encoding)
- Latin-1/ISO-8859-1 (extended ASCII)
## 🛠️ Solution
Replaced the simple `String::from_utf8_lossy()` call with intelligent
encoding detection via a new `bytes_to_string_smart()` function that
tries multiple encoding strategies:
1. **UTF-8** (fast path for valid UTF-8)
2. **Windows-1252** (handles Windows-specific characters in 0x80-0x9F
range)
3. **Latin-1** (fallback for extended ASCII)
4. **Lossy UTF-8** (final fallback, same as before)
## 📁 Changes
### New Files
- `codex-rs/core/src/text_encoding.rs` - Smart encoding detection module
- `codex-rs/core/tests/suite/text_encoding_fix.rs` - Integration tests
### Modified Files
- `codex-rs/core/src/lib.rs` - Added text_encoding module
- `codex-rs/core/src/exec.rs` - Updated StreamOutput::from_utf8_lossy()
- `codex-rs/core/tests/suite/mod.rs` - Registered new test module
## ✅ Testing
- **5 unit tests** covering UTF-8, Windows-1252, Latin-1, and fallback
scenarios
- **2 integration tests** simulating the exact Issue #6178 scenario
- **Demonstrates improvement** over the previous
`String::from_utf8_lossy()` approach
All tests pass:
```bash
cargo test -p codex-core text_encoding
cargo test -p codex-core test_shell_output_encoding_issue_6178
```
## 🎯 Impact
- ✅ **Eliminates garbled text** in VSCode shell preview for non-ASCII
content
- ✅ **Supports Windows/WSL environments** with proper encoding detection
- ✅ **Zero performance impact** for UTF-8 text (fast path)
- ✅ **Backward compatible** - UTF-8 content works exactly as before
- ✅ **Handles edge cases** with robust fallback mechanism
## 🧪 Test Scenarios
The fix has been tested with:
- Russian text ("пример")
- Windows-1252 quotation marks (""test")
- Latin-1 accented characters ("café")
- Mixed encoding content
- Invalid byte sequences (graceful fallback)
## 📋 Checklist
- [X] Addresses the reported issue
- [X] Includes comprehensive tests
- [X] Maintains backward compatibility
- [X] Follows project coding conventions
- [X] No breaking changes
---------
Co-authored-by: Josh McKinney <joshka@openai.com>
The `cap_sid` file contains the IDs of the two custom SIDs that the
Windows sandbox creates/manages to implement read-only and
workspace-write sandbox policies.
It previously lived in `<cwd>/.codex` which means that the sandbox could
write to it, which could degrade the efficacy of the sandbox. This
change moves it to `~/.codex/` (or wherever `CODEX_HOME` points to) so
that it is outside the workspace.
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.
## 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>
This allows `gh api` to work in the workspace-write sandbox w/ network
enabled. Without this we see e.g.
```
$ codex debug seatbelt --full-auto gh api repos/openai/codex/pulls --paginate -X GET -F state=all
Get "https://api.github.com/repos/openai/codex/pulls?per_page=100&state=all": tls: failed to verify certificate: x509: OSStatus -26276
```
- Added the new codex-windows-sandbox crate that builds both a library
entry point (run_windows_sandbox_capture) and a CLI executable to launch
commands inside a Windows restricted-token sandbox, including ACL
management, capability SID provisioning, network lockdown, and output
capture
(windows-sandbox-rs/src/lib.rs:167, windows-sandbox-rs/src/main.rs:54).
- Introduced the experimental WindowsSandbox feature flag and wiring so
Windows builds can opt into the sandbox:
SandboxType::WindowsRestrictedToken, the in-process execution path, and
platform sandbox selection now honor the flag (core/src/features.rs:47,
core/src/config.rs:1224, core/src/safety.rs:19,
core/src/sandboxing/mod.rs:69, core/src/exec.rs:79,
core/src/exec.rs:172).
- Updated workspace metadata to include the new crate and its
Windows-specific dependencies so the core crate can link against it
(codex-rs/
Cargo.toml:91, core/Cargo.toml:86).
- Added a PowerShell bootstrap script that installs the Windows
toolchain, required CLI utilities, and builds the workspace to ease
development
on the platform (scripts/setup-windows.ps1:1).
- Landed a Python smoke-test suite that exercises
read-only/workspace-write policies, ACL behavior, and network denial for
the Windows sandbox
binary (windows-sandbox-rs/sandbox_smoketests.py:1).
Currently, we change the tool description according to the sandbox
policy and approval policy. This breaks the cache when the user hits
`/approvals`. This PR does the following:
- Always use the shell with escalation parameter:
- removes `create_shell_tool_for_sandbox` and always uses unified tool
via `create_shell_tool`
- Reject the func call when the model uses escalation parameter when it
cannot.
When serializing to JSON, the existing solution created an enormous
array of ints, which is far more bytes on the wire than a base64-encoded
string would be.
This is a stopgap solution, but today, we are seeing the client get
flooded with events. Since we already truncate the output we send to the
model, it feels reasonable to limit how many deltas we send to the
client.
We want to send an aggregated output of stderr and stdout so we don't
have to aggregate it stderr+stdout as we lose order sometimes.
---------
Co-authored-by: Gabriel Peal <gpeal@users.noreply.github.com>
The high-order bit on this PR is that it makes it so `sandbox.rs` tests
both Mac and Linux, as we introduce a general
`spawn_command_under_sandbox()` function with platform-specific
implementations for testing.
An important, and interesting, discovery in porting the test to Linux is
that (for reasons cited in the code comments), `/dev/shm` has to be
added to `writable_roots` on Linux in order for `multiprocessing.Lock`
to work there. Granting write access to `/dev/shm` comes with some
degree of risk, so we do not make this the default for Codex CLI.
Piggybacking on top of #2317, this moves the
`python_multiprocessing_lock_works` test yet again, moving
`codex-rs/core/tests/sandbox.rs` to `codex-rs/exec/tests/sandbox.rs`
because in `codex-rs/exec/tests` we can use `cargo_bin()` like so:
```
let codex_linux_sandbox_exe = assert_cmd::cargo::cargo_bin("codex-exec");
```
which is necessary so we can use `codex_linux_sandbox_exe` and therefore
`spawn_command_under_linux_sandbox` in an integration test.
This also moves `spawn_command_under_linux_sandbox()` out of `exec.rs`
and into `landlock.rs`, which makes things more consistent with
`seatbelt.rs` in `codex-core`.
For reference, https://github.com/openai/codex/pull/1808 is the PR that
made the change to Seatbelt to get this test to pass on Mac.
This PR does two things because after I got deep into the first one I
started pulling on the thread to the second:
- Makes `ConversationManager` the place where all in-memory
conversations are created and stored. Previously, `MessageProcessor` in
the `codex-mcp-server` crate was doing this via its `session_map`, but
this is something that should be done in `codex-core`.
- It unwinds the `ctrl_c: tokio::sync::Notify` that was threaded
throughout our code. I think this made sense at one time, but now that
we handle Ctrl-C within the TUI and have a proper `Op::Interrupt` event,
I don't think this was quite right, so I removed it. For `codex exec`
and `codex proto`, we now use `tokio::signal::ctrl_c()` directly, but we
no longer make `Notify` a field of `Codex` or `CodexConversation`.
Changes of note:
- Adds the files `conversation_manager.rs` and `codex_conversation.rs`
to `codex-core`.
- `Codex` and `CodexSpawnOk` are no longer exported from `codex-core`:
other crates must use `CodexConversation` instead (which is created via
`ConversationManager`).
- `core/src/codex_wrapper.rs` has been deleted in favor of
`ConversationManager`.
- `ConversationManager::new_conversation()` returns `NewConversation`,
which is in line with the `new_conversation` tool we want to add to the
MCP server. Note `NewConversation` includes `SessionConfiguredEvent`, so
we eliminate checks in cases like `codex-rs/core/tests/client.rs` to
verify `SessionConfiguredEvent` is the first event because that is now
internal to `ConversationManager`.
- Quite a bit of code was deleted from
`codex-rs/mcp-server/src/message_processor.rs` since it no longer has to
manage multiple conversations itself: it goes through
`ConversationManager` instead.
- `core/tests/live_agent.rs` has been deleted because I had to update a
bunch of tests and all the tests in here were ignored, and I don't think
anyone ever ran them, so this was just technical debt, at this point.
- Removed `notify_on_sigint()` from `util.rs` (and in a follow-up, I
hope to refactor the blandly-named `util.rs` into more descriptive
files).
- In general, I started replacing local variables named `codex` as
`conversation`, where appropriate, though admittedly I didn't do it
through all the integration tests because that would have added a lot of
noise to this PR.
---
[//]: # (BEGIN SAPLING FOOTER)
Stack created with [Sapling](https://sapling-scm.com). Best reviewed
with [ReviewStack](https://reviewstack.dev/openai/codex/pull/2240).
* #2264
* #2263
* __->__ #2240
## Summary
A split-up PR of #1763 , stacked on top of a tools refactor #1858 to
make the change clearer. From the previous summary:
> Let's try something new: tell the model about the sandbox, and let it
decide when it will need to break the sandbox. Some local testing
suggests that it works pretty well with zero iteration on the prompt!
## Testing
- [x] Added unit tests
- [x] Tested locally and it appears to work smoothly!
## Summary
Escalating out of sandbox is (almost always) not going to fix
long-running commands timing out - therefore we should just pass the
failure back to the model instead of asking the user to re-run a command
that took a long time anyway.
## Testing
- [x] Ran locally with a timeout and confirmed this worked as expected
## Summary
Users frequently complain about re-approving commands that have failed
for non-sandbox reasons. We can't diagnose with complete accuracy which
errors happened because of a sandbox failure, but we can start to
eliminate some common simple cases.
This PR captures the most common case I've seen, which is a `command not
found` error.
## Testing
- [x] Added unit tests
- [x] Ran a few cases locally
## Summary
- stream command stdout as `ExecCommandStdout` events
- forward streamed stdout to clients and ignore in human output
processor
- adjust call sites for new streaming API
