## Why A parent turn can acknowledge a child agent's final reply while a model request is in flight. If the parent is interrupted before consuming the queued reply, turn cleanup can discard it. ## What changed Record queued agent messages in conversation history during interruption and flush the rollout before emitting `TurnAborted`, even when the interruption marker is disabled. ## Testing Add a regression test that interrupts a parent while a child's final reply is queued. The test checks that the reply is persisted when `TurnAborted` arrives and appears exactly once in the next model request. GitOrigin-RevId: 5bd828c4101942fa4fe7ba5040355a8244f02018
codex-core
This crate implements the business logic for Codex. It is designed to be used by the various Codex UIs written in Rust.
Wine-exec integration tests
On x86-64 Linux, run the shared suite against the Windows exec server with
bazel test //codex-rs/core:core-all-wine-exec-test.
Local execution targets the host OS, Docker targets Linux, and Wine exec targets Windows. Choose the skip macro by what the test depends on:
skip_if_target_windows!: Windows target behavior.skip_if_host_windows!: Windows host constraints.skip_if_remote!: Local-only test behavior.skip_if_no_remote_env!: Remote-only test behavior.skip_if_wine_exec!: Wine-specific runner debt.
Dependencies
Note that codex-core makes some assumptions about certain helper utilities being available in the environment. Currently, this support matrix is:
macOS
Expects /usr/bin/sandbox-exec to be present.
When using the workspace-write sandbox policy, the Seatbelt profile allows
writes under the configured writable roots while keeping .git (directory or
pointer file), the resolved gitdir: target, and .codex read-only.
Network access and filesystem read/write roots are controlled by
SandboxPolicy. Seatbelt consumes the resolved policy and enforces it.
Seatbelt also keeps the legacy default preferences read access
(user-preference-read) needed for cfprefs-backed macOS behavior.
Linux
Expects the binary containing codex-core to run the equivalent of codex sandbox when arg0 is codex-linux-sandbox. See the codex-arg0 crate for details.
The Linux sandbox accepts both legacy sandbox_mode configs and permission
profiles. Bubblewrap enforces filesystem restrictions and provides the network
namespace for managed proxy routes. A managed route selects bubblewrap even
when features.use_legacy_landlock is set. Without a managed route, the helper
rejects legacy Landlock if the policy restricts filesystem writes. With full
disk write access it can apply network restrictions in process.
The Linux sandbox helper prefers the first bwrap found on PATH outside the
current working directory whenever it is available. If bwrap is present but
too old to support --argv0, the helper keeps using system bubblewrap and
switches to a no---argv0 compatibility path for the inner re-exec. If
bwrap is missing, it falls back to the bundled codex-resources/bwrap
binary shipped with Codex and Codex surfaces a startup warning through its
normal notification path instead of printing directly from the sandbox helper.
Codex also surfaces a startup warning when bubblewrap cannot create user
namespaces. WSL2 uses the normal Linux bubblewrap path. WSL1 is not supported
for bubblewrap sandboxing because it cannot create the required user
namespaces, so Codex rejects sandboxed shell commands that would enter the
bubblewrap path before invoking bwrap.
Windows
Legacy SandboxPolicy / sandbox_mode configs are still supported on
Windows. Legacy read-only and workspace-write policies imply full
filesystem read access; exact readable roots are represented by split
filesystem policies instead.
The elevated Windows sandbox also supports:
- legacy
ReadOnlyandWorkspaceWritebehavior - split filesystem policies that need exact readable roots, exact writable roots, or extra read-only carveouts under writable roots
- backend-managed system read roots required for basic execution, such as
C:\Windows,C:\Program Files,C:\Program Files (x86), andC:\ProgramData, when a split filesystem policy requests platform defaults
The unelevated restricted-token backend still supports the legacy full-read
Windows model for legacy ReadOnly and WorkspaceWrite behavior. It also
supports a narrow split-filesystem subset: full-read split policies whose
writable roots still match the legacy WorkspaceWrite root set, but add extra
read-only carveouts under those writable roots.
New [permissions] / split filesystem policies remain supported on Windows
only when they can be enforced directly by the selected Windows backend or
round-trip through the legacy SandboxPolicy model without changing semantics.
Policies that would require direct explicit unreadable carveouts (none) or
reopened writable descendants under read-only carveouts still fail closed
instead of running with weaker enforcement.
All Platforms
Expects the binary containing codex-core to simulate the virtual
apply_patch CLI when arg1 is --codex-run-as-apply-patch. See the
codex-arg0 crate for details.
Runtime authentication (experimental)
Runtime authentication is off by default. To enable it in config.toml:
[features.runtime_agent_instances]
enabled = true
mode = "dual_header"
dual_header sends the existing actor credential and a fresh agent assertion.
assertion_only sends only the assertion. The latter requires backend support;
registration failures do not fall back to credential-only requests. Non-OpenAI
providers and third-party MCP authentication keep their existing behavior.
The same table in requirements.toml pins enablement and mode. A conflicting
explicit mode is rejected. The older boolean runtime_agent_instances = true
still works and uses the default mode (dual_header) unless requirements pin it.
Deliver typed requirements only to client versions that understand them.
The harness owns the runtime and its memory-only signing key. Ordinary token
refreshes reuse it; a different credential source or actor context requires a new
runtime. When runtime authentication is selected, legacy AgentIdentity
credentials migrate to saved OAuth credentials for the same actor and remove the
persisted signing key. If no matching actor credential exists, a supported
sign-in is required. Ordinary credential loading does not perform this migration.
Cloud-config bundle and requirements fetches retain their endpoint-restricted bootstrap authentication, including refreshes after runtime initialization.