## Why
The `spawn_agent` model override guidance is uncapped and bloating
context. We need to trim down each entry and cap total entries.
picked 5 as cap, we can change
## What changed
- Cap the model override summaries shown in `spawn_agent` to the first 5
picker-visible models, preserving the existing priority ordering from
the models manager.
- Condense each rendered entry to the actionable pieces the model needs:
- use the model slug as the label
- render compact reasoning effort lists with the default marked inline
- render only service tier IDs, and omit the clause when no tiers are
available
- Update coverage so the compact formatter shape and the top-5 cap are
exercised, and keep the end-to-end request assertion aligned with real
model metadata.
## Example
Before:
`- gpt-5.4 ('gpt-5.4\'): Strong model for everyday coding. Default
reasoning effort: medium. Supported reasoning efforts: low (Fast
responses with lighter reasoning), medium (Balances speed and reasoning
depth for everyday tasks), high (Greater reasoning depth for complex
problems), xhigh (Extra high reasoning depth for complex problems).
Supported service tiers: priority (Fast: 1.5x speed, increased usage).`
After:
`- 'gpt-5.4': Strong model for everyday coding. Reasoning efforts: low,
medium (default), high, xhigh. Service tiers: priority.`
## Why
`SandboxPolicy` is now a legacy compatibility shape, but several tests
still built a `SandboxPolicy` only to immediately convert it into
`PermissionProfile` for APIs that already accept canonical runtime
permissions. Those detours make it harder to audit where legacy sandbox
policy is still required, because boundary-only usages are mixed
together with ordinary test setup.
## What Changed
- Updated tests in `codex-core`, `codex-exec`, `codex-analytics`, and
`codex-config` to construct `PermissionProfile` values directly when the
code under test takes a permission profile.
- Changed exec-policy, request-permissions, session, and sandbox test
helpers to pass `PermissionProfile` through instead of converting from
`SandboxPolicy` internally.
- Left `SandboxPolicy` in place where tests are explicitly exercising
legacy compatibility or request/response boundaries.
## Test Plan
- `cargo test -p codex-analytics -p codex-config`
- `cargo test -p codex-core --lib safety::tests`
- `cargo test -p codex-core --lib exec_policy::tests::`
- `cargo test -p codex-core --lib exec::tests`
- `cargo test -p codex-core --lib guardian_review_session_config`
- `cargo test -p codex-core --lib tools::network_approval::tests`
- `cargo test -p codex-core --lib
tools::runtimes::shell::unix_escalation::tests`
- `cargo test -p codex-core --lib managed_network`
- `cargo test -p codex-core --test all request_permissions::`
- `cargo test -p codex-exec sandbox`
---
[//]: # (BEGIN SAPLING FOOTER)
Stack created with [Sapling](https://sapling-scm.com). Best reviewed
with [ReviewStack](https://reviewstack.dev/openai/codex/pull/23030).
* #23036
* __->__ #23030
## Why
Goal completion follow-up turns currently receive a preformatted English
usage sentence such as `time used: 2586 seconds`. That nudges the model
to echo an awkward raw seconds count in the final reply, even though the
tool result already exposes structured usage fields like
`goal.timeUsedSeconds`, `goal.tokensUsed`, and `goal.tokenBudget`.
## What changed
- Replace the preformatted completion usage sentence with guidance to
read the structured goal fields from the tool result.
- Preserve token-budget reporting while allowing the model to phrase
elapsed time in a concise, human-friendly way that fits the response
language.
- Update core coverage for both the generated completion guidance and
the session flow that forwards it back to the model.
## Verification
Previously, it would have output a final message indicating that it
"worked for 303 seconds". Now it shows the following:
<img width="286" height="35" alt="image"
src="https://github.com/user-attachments/assets/d7011880-9449-46a7-856f-4e50ae00eb45"
/>
## Summary
- Add optional image detail to user image inputs across core, app-server
v2, thread history/event mapping, and the generated app-server
schemas/types.
- Preserve requested detail when serializing Responses image inputs:
omitted detail stays on the existing `high` default, while explicit
`original` keeps local images on the original-resolution path.
- Support `high`/`original` consistently for tool image outputs,
including MCP `codex/imageDetail`, code-mode image helpers, and
`view_image`.
## Summary
- move tool_user_shell_type out of the old tools::spec module and call
it from tools directly
- attach the remaining spec planning model tests under spec_plan
- delete core/src/tools/spec.rs
## Tests
- just fmt
- cargo test -p codex-core tools::spec_plan
Note: a broader cargo test -p codex-core run on the earlier PR-head
worktree still hit the pre-existing stack overflow in
agent::control::tests::spawn_agent_fork_last_n_turns_keeps_only_recent_turns.
## Why
The tool runtime path still had a typed output associated type on
`ToolExecutor`, plus a core-only `RegisteredTool` adapter and
extension-only executor aliases. That made every new shared tool runtime
carry extra adapter plumbing before it could participate in core
dispatch, extension tools, hook payloads, telemetry, and model-visible
spec generation.
This PR moves output erasure to the shared executor boundary so core and
extension tools can use the same execution contract directly.
## What Changed
- Changed `codex_tools::ToolExecutor` to return `Box<dyn ToolOutput>`
instead of an associated `Output` type.
- Removed the extension-specific `ExtensionToolExecutor` /
`ExtensionToolOutput` aliases and exposed `ToolExecutor<ToolCall>` plus
`ToolOutput` through `codex-extension-api`.
- Reworked core tool registration around `CoreToolRuntime` and
`ToolRegistry::from_tools`, removing the extra `RegisteredTool` /
`ToolRegistryBuilder` layer.
- Consolidated model-visible spec planning and registry construction in
`core/src/tools/spec_plan.rs`, including deferred tool search and
code-mode-only filtering.
- Added `ToolOutput` helpers for post-tool-use hook ids and inputs so
MCP, unified exec, extension, and other boxed outputs preserve the same
hook payload behavior.
- Updated core handlers, memories tools, and the related
registry/spec/router tests to use the simplified contract.
## Test Coverage
- Updated coverage for tool spec planning, registry lookup, deferred
tool search registration, extension tool routing, post-tool-use hook
payloads, dispatch tracing, guardian output extraction, and memories
extension tool execution.
## Why
This PR is the invariant-cleanup layer that follows the workspace-roots
base merged in [#22610](https://github.com/openai/codex/pull/22610).
#22610 adds `[permissions.<id>.workspace_roots]` and keeps runtime
workspace roots separate from the raw permission profile, but its
in-memory representation is intentionally transitional: `Permissions`
still carries the selected profile identity next to a constrained
`PermissionProfile`. That makes APIs such as
`set_constrained_permission_profile_with_active_profile()` fragile
because the id and value only mean the right thing when every caller
keeps them in sync.
This PR introduces a single resolved profile state so profile identity,
`extends`, the profile value, and profile-declared workspace roots
travel together. The next PR,
[#22611](https://github.com/openai/codex/pull/22611), builds on this by
changing the app-server turn API to select permission profiles by id
plus runtime workspace roots.
## Stack Context
- #22610, now merged: adds profile-declared `workspace_roots`, runtime
workspace roots, and `:workspace_roots` materialization.
- This PR: replaces the parallel active-profile/profile-value fields
with `PermissionProfileState`.
- #22611: switches app-server turn updates toward profile ids plus
runtime workspace roots.
- #22612: updates TUI/exec summaries to show the effective workspace
roots.
Keeping this separate from #22611 is deliberate: reviewers can validate
the internal state invariant before reviewing the app-server protocol
migration.
## What Changed
- Added `ResolvedPermissionProfile::{Legacy, BuiltIn, Named}` and
`PermissionProfileState`.
- Typed built-in profile ids with `BuiltInPermissionProfileId`.
- Moved selected profile identity and profile-declared workspace roots
into the resolved state.
- Replaced `Permissions` parallel profile fields with one
`permission_profile_state`.
- Removed `set_constrained_permission_profile_with_active_profile()`
from session sync paths.
- Kept trusted session replay/`SessionConfigured` compatibility through
explicit session snapshot helpers.
- Updated session configuration, MCP initialization, app-server, exec,
TUI, and guardian call sites to consume `&PermissionProfile` directly.
## Review Guide
Start with `codex-rs/core/src/config/resolved_permission_profile.rs`; it
is the new invariant boundary. Then review
`codex-rs/core/src/config/mod.rs` to see how config loading records
active profile identity and profile workspace roots. The remaining
call-site changes are mostly mechanical fallout from
`Permissions::permission_profile()` returning `&PermissionProfile`
instead of `&Constrained<PermissionProfile>`.
## Verification
The existing config/session coverage now constructs and asserts through
`PermissionProfileState`. The workspace-root config test also asserts
that profile-declared roots are preserved in the resolved state, which
is the behavior #22611 relies on when runtime roots become mutable
through the app-server API.
---
[//]: # (BEGIN SAPLING FOOTER)
Stack created with [Sapling](https://sapling-scm.com). Best reviewed
with [ReviewStack](https://reviewstack.dev/openai/codex/pull/22683).
* #22612
* #22611
* __->__ #22683
## Why
This is the configuration/model half of the alternative permissions
migration we discussed as a comparison point for
[#22401](https://github.com/openai/codex/pull/22401) and
[#22402](https://github.com/openai/codex/pull/22402).
The old `workspace-write` model mixes three concerns that we want to
keep separate:
- reusable profile rules that should stay immutable once selected
- user/runtime workspace roots from `cwd`, `--add-dir`, and legacy
workspace-write config
- internal Codex writable roots such as memories, which should not be
shown as user workspace roots
This PR gives permission profiles first-class `workspace_roots` so users
can opt multiple repositories into the same `:workspace_roots` rules
without using broad absolute-path write grants. It also starts
separating the raw selected profile from the effective runtime profile
by making `Permissions` expose explicit accessors instead of public
mutable fields.
A representative `config.toml` looks like this:
```toml
default_permissions = "dev"
[permissions.dev.workspace_roots]
"~/code/openai" = true
"~/code/developers-website" = true
[permissions.dev.filesystem.":workspace_roots"]
"." = "write"
".codex" = "read"
".git" = "read"
".vscode" = "read"
```
If Codex starts in `~/code/codex` with that profile selected, the
effective workspace-root set becomes:
- `~/code/codex` from the runtime `cwd`
- `~/code/openai` from the profile
- `~/code/developers-website` from the profile
The `:workspace_roots` rules are materialized across each root, so
`.git`, `.codex`, and `.vscode` stay scoped the same way everywhere.
Runtime additions such as `--add-dir` can still layer on later stack
entries without mutating the selected profile.
## Stack Shape
This PR intentionally stops before the profile-identity cleanup in
[#22683](https://github.com/openai/codex/pull/22683) so the base review
stays focused on config loading, workspace-root materialization, and
compatibility with legacy `workspace-write`.
The representation in this PR is therefore transitional: `Permissions`
carries enough state to distinguish the raw constrained profile from the
effective runtime profile, and there are still call sites that must keep
the active profile identity and constrained profile value in sync. The
follow-up PR replaces that with a single resolved profile state
(`ResolvedPermissionProfile` / `PermissionProfileState`) that keeps the
profile id, immutable `PermissionProfile`, and profile-declared
workspace roots together. That follow-up removes APIs such as
`set_constrained_permission_profile_with_active_profile()` where
separate arguments could drift out of sync.
Downstream PRs then build on this base to switch app-server turn updates
to profile ids plus runtime workspace roots and to finish the
user-visible summary behavior. Reviewers should judge this PR as the
workspace-roots foundation, not as the final in-memory shape of selected
permission profiles.
## Review Guide
Suggested review order:
1. Start with `codex-rs/core/src/config/mod.rs`.
This is the main shape change in the base slice. `Permissions` now
stores a private raw `Constrained<PermissionProfile>` plus runtime
`workspace_roots`. Callers use `permission_profile()` when they need the
raw constrained value and `effective_permission_profile()` when they
need a materialized runtime profile. As noted above,
[#22683](https://github.com/openai/codex/pull/22683) replaces this
transitional shape with a resolved profile state that keeps identity and
profile data together.
2. Review `codex-rs/config/src/permissions_toml.rs` and
`codex-rs/core/src/config/permissions.rs`.
These add `[permissions.<id>.workspace_roots]`, resolve enabled entries
relative to the policy cwd, and keep `:workspace_roots` deny-read glob
patterns symbolic until the actual roots are known.
3. Review `codex-rs/protocol/src/permissions.rs` and
`codex-rs/protocol/src/models.rs`.
These add the policy/profile materialization helpers that expand exact
`:workspace_roots` entries and scoped deny-read globs over every
workspace root. This is also where `ActivePermissionProfileModification`
is removed from the core model.
4. Review the legacy bridge in
`Config::load_from_base_config_with_overrides` and
`Config::set_legacy_sandbox_policy`.
This is where legacy `workspace-write` roots become runtime workspace
roots, while Codex internal writable roots stay internal and do not
appear as user-facing workspace roots.
5. Then skim downstream call sites.
The interesting pattern is raw-vs-effective access: state/proxy/bwrap
paths keep the raw constrained profile, while execution, summaries, and
user-visible status use the effective profile and workspace-root list.
## What Changed
- added `[permissions.<id>.workspace_roots]` to the config model and
schema
- added runtime `workspace_roots` state to `Config`/`Permissions` and
`ConfigOverrides`
- made `Permissions` profile fields private and replaced direct mutation
with accessors/setters
- added `PermissionProfile` and `FileSystemSandboxPolicy` helpers for
materializing `:workspace_roots` exact paths and deny-read globs across
all roots
- moved legacy additional writable roots into runtime workspace-root
state instead of active profile modifications
- removed `ActivePermissionProfileModification` and its app-server
protocol/schema export
- updated sandbox/status summary paths so internal writable roots are
not reported as user workspace roots
## Verification Strategy
The targeted tests cover the behavior at the layers where regressions
are most likely:
- `codex-rs/core/src/config/config_tests.rs` verifies config loading,
legacy workspace-root seeding, effective profile materialization, and
memory-root handling.
- `codex-rs/core/src/config/permissions_tests.rs` verifies profile
`workspace_roots` parsing and `:workspace_roots` scoped/glob
compilation.
- `codex-rs/protocol/src/permissions.rs` unit tests verify exact and
glob materialization over multiple workspace roots.
- `codex-rs/tui/src/status/tests.rs` and
`codex-rs/utils/sandbox-summary/src/sandbox_summary.rs` verify the
user-facing summaries show effective workspace roots and hide internal
writes.
I also ran `cargo check --tests` locally after the latest stack refresh
to catch cross-crate API breakage from the private-field/accessor
changes.
---
[//]: # (BEGIN SAPLING FOOTER)
Stack created with [Sapling](https://sapling-scm.com). Best reviewed
with [ReviewStack](https://reviewstack.dev/openai/codex/pull/22610).
* #22612
* #22611
* #22683
* __->__ #22610
## Summary
Removes the feature since this is effectively on by default in all cases
where we should use it, or can be configured via models.json.
## Testing
- [x] unit tests pass
## Why
`codex_tools::ToolExecutor` keeps a tool spec attached to its runtime
handler, but extension tools still carried a parallel
`ExtensionToolFuture` / `ExtensionToolExecutor` shape. That made
extension-owned tools look different from host tools even though
routing, registration, and execution need the same abstraction.
This PR makes the shared executor contract directly async and lets
extension tools implement it too, so host tools and extension tools can
move through the same registration path.
## What changed
- Changed `ToolExecutor::handle` to an `async fn` using `async-trait`,
and updated built-in tool handlers to implement the async trait
directly.
- Replaced the bespoke `ExtensionToolFuture` contract with a marker
`ExtensionToolExecutor` over `ToolExecutor<ToolCall, Output =
JsonToolOutput>`, re-exporting `ToolExecutor` from
`codex-extension-api`.
- Updated the memories extension tools to implement the shared executor
trait.
- Split tool-router construction into collected executors plus hosted
model specs, keeping hosted tools like web search and image generation
separate from executable handlers.
- Updated spec/router tests and extension-tool stubs for the new
executor shape.
## Verification
- Not run locally.
## Why
`multi_agent_v2` already allowed configuring the minimum `wait_agent`
timeout, but the default timeout and upper bound were still hard-coded.
That made it hard to tune waits for subagent mailbox activity in
sessions that need either faster wakeups or longer waits, and it meant
the model-visible `wait_agent` schema could not fully reflect the
resolved runtime limits.
## What Changed
- Added `features.multi_agent_v2.max_wait_timeout_ms` and
`features.multi_agent_v2.default_wait_timeout_ms` alongside the existing
`min_wait_timeout_ms` setting.
- Validated all three timeouts in config as `0..=3_600_000`, with
`min_wait_timeout_ms <= default_wait_timeout_ms <= max_wait_timeout_ms`.
- Thread and review session tool config now passes the resolved
min/default/max values into the `wait_agent` tool schema.
- `wait_agent` now uses the configured default when `timeout_ms` is
omitted and rejects explicit values outside the configured min/max range
instead of silently clamping them.
- Updated the generated config schema and config-lock test coverage for
the new fields.
## Why
On Windows, elevated sandboxed commands run under a dedicated sandbox
account while `HOME` / `USERPROFILE` can still point at the real user's
profile directory. For PowerShell login shells, that combination can
make the sandbox account try to load the real user's PowerShell profile
script. If the sandbox account's execution policy differs from the real
user's policy, startup can emit profile-loading errors before the
requested command runs.
For this backend, loading the profile is not a faithful user login
shell: it is cross-account profile execution. Treating these PowerShell
invocations as non-login shells avoids that invalid startup path.
## Why This Happens Late
The normal `login` decision is resolved when shell argv is created, but
that point is too early to make this Windows sandbox-specific decision.
At argv creation time we do not yet know the actual sandbox attempt that
will run the command. A turn can include sandboxed and unsandboxed
attempts, and a broad turn-level override would also affect Full Access
commands where the user's profile should remain available.
Instead, this change carries the selected `ShellType` alongside the argv
and applies the `-NoProfile` adjustment in the shell runtimes once the
`SandboxAttempt` is known. That keeps the override scoped to actual
`WindowsRestrictedToken` attempts with `WindowsSandboxLevel::Elevated`.
The runtime uses the selected shell metadata rather than re-detecting
PowerShell from argv. That avoids brittle parsing and covers PowerShell
invocation shapes such as `-EncodedCommand`.
## What Changed
- Carry selected shell metadata through `exec_command` / unified exec
requests and shell tool requests.
- Insert `-NoProfile` for PowerShell commands only when the runtime is
about to execute a sandboxed elevated Windows attempt.
- Add focused unit coverage for elevated Windows PowerShell,
`-EncodedCommand`, existing `-NoProfile`, legacy restricted-token
attempts, unsandboxed attempts, and non-PowerShell commands.
## Verification
- `cargo test -p codex-core disable_powershell_profile_tests`
- `cargo test -p codex-core test_get_command`
- `cargo clippy --fix --tests --allow-dirty --allow-no-vcs -p
codex-core`
A full `cargo test -p codex-core` run was also attempted during
development, but it still hit an unrelated stack overflow in
`agent::control` tests before reaching this area.
## Why
`TurnContext::cwd` is deprecated in favor of resolving paths from the
selected turn environment cwd. A few filesystem-oriented paths were
still constructing sandbox context from the legacy cwd and then mutating
it afterward, or resolving local file paths through the deprecated
helper.
## What changed
- Make `TurnContext::file_system_sandbox_context` take the trusted cwd
explicitly.
- Pass the selected turn environment cwd directly from `apply_patch` and
`view_image` call sites.
- Restrict `spawn_agents_on_csv` to exactly one local environment and
resolve input/output CSV paths from that local environment cwd.
- Remove a redundant test setup assignment that only synchronized
deprecated `TurnContext::cwd` with a replaced config.
## Validation
- `cargo test -p codex-core view_image`
- `cargo test -p codex-core
maybe_persist_mcp_tool_approval_writes_project_config_for_project_server`
- `cargo test -p codex-core parse_csv_supports_quotes_and_commas`
- `git diff --check`
## Why
`TurnContext::cwd` and `TurnContext::resolve_path` are being phased out
in favor of using the selected turn environment cwd directly.
Deprecating both APIs makes any new direct dependency visible while
preserving the existing migration path for current callers.
## What Changed
- Marked `TurnContext::cwd` and `TurnContext::resolve_path` as
deprecated with guidance to use the selected turn environment cwd
instead.
- Added exact `#[allow(deprecated)]` suppressions at each existing
direct usage site, including tests, rather than adding crate-wide
suppression.
- Kept the change behavior-preserving: current cwd reads, writes, and
path resolution continue to use the same values.
## Verification
- `just fmt`
- `cargo check -p codex-core`
- `cargo check -p codex-core --tests`
- `git diff --check`
## Why
`code_mode_only` filters code-mode nested tools out of the top-level
tool list. For multi-agent v2, we need a rollout shape where the
collaboration tools remain callable as normal model tools without also
being embedded into the code-mode `exec` tool declaration.
Related to this:
https://openai-corpws.slack.com/archives/C0AQLHB4U75/p1778660267922549
## What Changed
- Adds `features.multi_agent_v2.non_code_mode_only`, including config
resolution, profile override handling, and generated schema coverage.
- Introduces `ToolExposure::DirectModelOnly` so a tool can be included
in the initial model-visible list while staying out of the nested
code-mode tool surface.
- Applies that exposure to the multi-agent v2 tools when the new flag is
set: `spawn_agent`, `send_message`, `followup_task`, `wait_agent`,
`close_agent`, and `list_agents`.
- Updates code-mode-only filtering so direct-model-only tools remain
visible while ordinary nested code-mode tools are still hidden.
## Verification
- Added config parsing/profile tests for `non_code_mode_only`.
- Added tool spec coverage for the code-mode-only multi-agent v2
exposure behavior.
## Why
Recent session history showed no active use of the raw `shell`,
`local_shell`, or `container.exec` execution surfaces. Keeping those
handlers/specs wired into core leaves duplicate shell execution paths
alongside the supported `shell_command` and unified exec tools.
## What changed
- Removed the raw `shell` handler/spec and its `ShellToolCallParams`
protocol helper.
- Removed the legacy `local_shell` and `container.exec` handler/spec
plumbing while preserving persisted-history compatibility for old
response items.
- Normalized model/config `default` and `local` shell selections to
`shell_command`.
- Pruned tests that exercised removed raw-shell/local-shell/apply-patch
variants and kept coverage on `shell_command`, unified exec, and
freeform `apply_patch`.
## Verification
- `git diff --check`
- `cargo test -p codex-protocol`
- `cargo test -p codex-tools`
- `cargo test -p codex-core tools::handlers::shell`
- `cargo test -p codex-core tools::spec`
- `cargo test -p codex-core tools::router`
- `cargo test -p codex-core
active_call_preserves_triggering_command_context`
- `cargo test -p codex-core guardian_tests`
- `cargo test -p codex-core --test all shell_serialization`
- `cargo test -p codex-core --test all apply_patch_cli`
- `cargo test -p codex-core --test all shell_command_`
- `cargo test -p codex-core --test all local_shell`
- `cargo test -p codex-core --test all otel::`
- `cargo test -p codex-core --test all hooks::`
- `just fix -p codex-core`
- `just fix -p codex-tools`
## Why
Deferred tools were tracked with separate side-channel filtering after
tool specs had already been assembled. That made the registry
responsible for executing tools while the router/spec planner separately
decided whether those same tools should be exposed to the model up
front.
This PR makes exposure part of the tool handler contract so direct
versus deferred availability travels with the executable tool
registration.
Next step will be to simplify registration
## What Changed
- Adds `ToolExposure` to `codex-tools` and exposes it through
`ToolExecutor`, defaulting tools to `Direct`.
- Teaches dynamic tools and MCP handlers to mark deferred tools as
`Deferred` at construction time.
- Renames the registry object-safe wrapper from `AnyToolHandler` to
`RegisteredTool` and uses `ToolExposure` when deciding whether to
include a handler's spec in the initial model-visible tool list.
- Refactors tool spec planning to derive direct specs and deferred
search entries from registered handlers, removing the router's
special-case deferred dynamic tool filtering.
## Verification
- Not run.
## Why
Spawned agents can already override `model` and `reasoning_effort`, but
they have no equivalent way to opt into a model-supported service tier.
That makes it impossible to preserve or intentionally select tiered
execution behavior when delegating work to a sub-agent, even though the
model catalog already advertises supported `service_tiers`.
## What changed
- Add optional `service_tier` to both legacy and `MultiAgentV2`
`spawn_agent` tool inputs.
- Show each picker-visible model's supported service tier ids and
descriptions in the `spawn_agent` tool guidance.
- Resolve service tier selection after the child agent's effective model
is known.
- Inherit the parent tier when omitted and still supported by the final
child model; otherwise clear it.
- Reject explicit unsupported tier requests with a model-facing error.
- Keep explicit `service_tier` usable on full-history forks, while still
honoring the existing model/reasoning fork restrictions.
- Hide `service_tier` alongside other spawn metadata when
`hide_spawn_agent_metadata` is enabled.
## Verification
Added focused coverage for:
- v1/v2 `spawn_agent` schema exposure for `service_tier`
- tier descriptions in spawn guidance
- hidden-metadata suppression
- explicit supported tier selection
- explicit unknown and unsupported tier rejection
- inherited tier preservation or clearing based on child-model support
- full-history fork acceptance for explicit service tiers in both v1 and
v2
Local Rust tests were not run in this workspace per repo guidance; the
new coverage is included for CI.
## Why
Extension tools were split across two public runtime contracts:
`codex-tool-api` exposed `ToolBundle` plus its own call/spec/error
types, while core native tools used `codex_tools::ToolExecutor`. That
made contributed tool specs and execution behavior easy to drift apart
and added another crate boundary for what should be one executable-tool
seam.
This PR makes `ToolExecutor` the single runtime contract and keeps
extension-specific pinning in `codex-extension-api`.
## Remaining todo
https://github.com/openai/codex/pull/22369/changes#diff-b935ea8245c3ce568a30cff660175fa6390b66b872ae409e1e2e965738250741R5
Either generic `Invocation` or sub-extract the `ToolCall` and clean
`ToolInvocation`
## What changed
- Removed the `codex-tool-api` workspace crate and its dependencies from
core and `codex-extension-api`.
- Made `codex_tools::ToolExecutor` object-safe with `async_trait` so
extension contributors can return a dyn executor.
- Added the extension-facing aliases under
`ext/extension-api/src/contributors/tools.rs`, including
`ExtensionToolExecutor = dyn ToolExecutor<ToolCall, Output =
ExtensionToolOutput>`.
- Changed `ToolContributor::tools` to return extension executors
directly instead of `ToolBundle`s.
- Updated core’s extension tool handler/registry/router path to adapt
those extension executors into the existing native `ToolInvocation`
runtime path.
- Added focused coverage for extension tools being registered,
model-visible, dispatchable, and not replacing built-in tools.
## Verification
- `cargo test -p codex-tools`
- `cargo test -p codex-extension-api`
## Why
Codex still models model-visible tools and executable behavior largely
inside `codex-core`, which makes it harder to evolve the tool system
toward a single reusable abstraction for built-ins, MCP-backed tools,
dynamic tools, and later tools injected from outside core.
This PR takes the next incremental step in that direction by moving the
common execution-facing pieces out of core and separating them from
core-only orchestration. The intent is to let shared tool abstractions
improve in one place, while `codex-core` keeps the parts that are still
inherently host-specific today, such as `ToolInvocation`, dispatch
wiring, and hook integration.
This PR is mostly moving things around. The only interesting piece is
this abstraction:
https://github.com/openai/codex/pull/22359/changes#diff-81af519002548ba51ed102bdaaf77e081d40a1e73a6e5f9b104bbbc96a6f1b3dR13
## What changed
- Added `codex_tools::ToolExecutor<Invocation>` as the shared execution
trait for model-visible tools.
- Moved the reusable execution support types from `codex-core` into
`codex-tools`:
- `FunctionCallError`
- `ToolPayload`
- `ToolOutput`
- Refactored core tool implementations so that execution behavior lives
on `ToolExecutor<ToolInvocation>`, while `ToolHandler` remains the
core-local extension point for hook payloads, telemetry tags, diff
consumers, and other orchestration concerns.
- Kept the registry and dispatch flow behaviorally unchanged while
making the shared/extracted boundary explicit across built-in, MCP,
dynamic, extension-backed, shell, and multi-agent tool handlers.
## Verification
- `cargo test -p codex-tools`
- `just fix -p codex-tools`
- `just fix -p codex-core`
- `cargo test -p codex-core` progressed through the updated tool
surfaces and then hit the existing unrelated multi-agent stack overflow
in
`tools::handlers::multi_agents::tests::tool_handlers_cascade_close_and_resume_and_keep_explicitly_closed_subtrees_closed`.
## Summary
- move the `view_image` sandbox filesystem-read unit test onto a
temporary cwd
- keep the turn cwd and selected turn environment cwd aligned inside the
test
- avoid leaving `core/image.png` behind in the repo checkout after the
test runs
## Root cause
The test wrote `image.png` beneath `turn.cwd`, and the shared session
test helper defaults that cwd to the current repo directory when no
override is provided.
## Validation
- `just fmt`
- `cargo test -p codex-core
tools::handlers::view_image::tests::handle_passes_sandbox_context_for_local_filesystem_reads`
## Why
`UnavailableDummyTools` kept synthetic placeholder tools alive for
historical tool calls whose backing MCP tool was no longer available.
That path adds stale model-visible tool specs and special routing at the
point where unavailable MCP calls should use ordinary current-tool
handling. This removes the runtime backfill instead of preserving a
second compatibility lane.
## Is it safe to remove?
The unavailable tools were added in #17853 after a CS issue when a
previously-called MCP tool failed to load and was omitted from the CS
spec. Now that we have tool search, I think this is resolved:
- API merges tools from previous TST output into effective tool set so
theyre always in CS spec
- if an MCP tool surfaced by TST later becomes unavailable, the model
can still call it and it will just return model-visible error
- both TST output and function call output are dropped on compaction so
model will not remember old calls to MCP post compaction
## What changed
- Delete unavailable-tool collection, placeholder handler, router/spec
plumbing, and obsolete placeholder coverage.
- Keep `features.unavailable_dummy_tools` as a removed no-op feature
tombstone so existing configs still parse cleanly.
- Add an integration-style `tool_search` regression test showing that a
deferred MCP tool surfaced through `tool_search` still routes through
MCP and returns a model-visible tool-call error rather than `unsupported
call`.
## Verification
- `cargo test -p codex-core tool_search`
## Why
This builds on the handler-owned spec refactor by moving deferred
tool-search metadata to the same handlers that already own tool specs.
The registry builder no longer needs a separate prebuilt
`tool_search_entries` path; it can collect searchable entries from
deferred handlers directly.
## What changed
- Added `search_info()` to tool handlers and implemented it for MCP and
dynamic handlers.
- Reused handler `spec()` output when constructing tool-search entries,
adapting it into the deferred `LoadableToolSpec` shape expected by
`tool_search`.
- Simplified `build_tool_registry_builder(...)` so `tool_search`
registration is based on deferred handlers with search info.
- Removed the old standalone search-entry builders and now-unused
`codex-tools` discovery helper exports.
## Verification
- `cargo test -p codex-core tools::handlers::tool_search::tests:: --
--nocapture`
- `cargo test -p codex-core tools::spec_plan::tests::search_tool --
--nocapture`
- `cargo test -p codex-core tools::spec::tests:: -- --nocapture`
- `cargo test -p codex-core tools::spec_plan::tests:: -- --nocapture`
- `cargo test -p codex-tools`
- `just fix -p codex-core`
- `just fix -p codex-tools`
## Why
Code mode already builds the merged nested `ToolSpec`s that feed the
`exec` prompt. Keeping a separate `tool_namespaces` map in the planning
path duplicated that metadata and left extra wrapper plumbing in
`spec.rs`.
## What changed
- derive code-mode namespace descriptions from the merged
`ToolSpec::Namespace` entries before building the code-mode handlers
- extract `build_code_mode_handlers(...)` so the code-mode-specific
planning stays in one place
- remove `tool_namespaces` from `ToolRegistryBuildParams`
- delete the now-unused `McpToolPlanInputs` wrapper and related test
helper plumbing
## Testing
- `cargo test -p codex-core spec_plan`
## Why
hook semantics treat `session_id` as shared across a root session and
its subagents. Codex hooks were still emitting the current thread ID,
which made spawned agents look like independent sessions and made it
harder for hook integrations to correlate work across a root thread and
its spawned helpers
This change makes hooks use Codex's existing shared session identity so
hook `session_id` matches the root-thread session across spawned
subagents.
## What Changed
- switch hook payloads to use the existing shared session identity from
core instead of the current thread ID
- cover all hook surfaces that expose `session_id`, including
`SessionStart`, tool hooks, compact hooks, prompt-submit hooks, stop
hooks, and legacy after-agent dispatch
## Why
`tool_search` still carries the server-specific result-cap path added in
#17684 for `computer-use`: when the model omitted `limit`, a matching
result expanded the search to 20 and then `limit_results_by_bucket`
applied per-bucket caps. That makes default result handling depend on a
one-off server exception instead of the single
`TOOL_SEARCH_DEFAULT_LIMIT` path.
This PR removes that custom branch so omitted `limit` values use the
ordinary global default consistently. The implementation being retired
is the pre-change bucketed search path in
[`tool_search.rs`](https://github.com/openai/codex/blob/5e3ee5eddfa5333f2e0b011880abf0cbf92bd295/codex-rs/core/src/tools/handlers/tool_search.rs#L121-L190).
## What changed
- Collapse `ToolSearchHandler::search` back to one BM25 search with the
resolved limit.
- Remove `limit_results_by_bucket`, the `computer-use` constants, and
the omitted-limit plumbing that only existed for the override.
- Drop dead `ToolSearchEntry::limit_bucket` metadata from deferred MCP
and dynamic search entries.
- Remove tests and helpers that only asserted the deleted override
behavior.
- Add direct handler-level unit coverage for omitted/default and
explicit `tool_search` result limits.
## Validation
- `cargo test -p codex-core tool_search`
- The matching unit tests passed, including the new omitted/default and
explicit result-limit coverage.
- The broader `--test all` search-tool fixture phase then failed before
sending mocked response requests in
`tool_search_indexes_only_enabled_non_app_mcp_tools` and
`tool_search_uses_non_app_mcp_server_instructions_as_namespace_description`.
- `cargo test -p codex-core`
- The touched tool-search coverage passed before the run later aborted
in
`tools::handlers::multi_agents::tests::tool_handlers_cascade_close_and_resume_and_keep_explicitly_closed_subtrees_closed`
with a stack overflow.
## Why
This is the base PR in the split stack for the permissions migration. It
isolates stack-safety work that had been mixed into the larger
permissions PR, so reviewers can evaluate the async-future changes
separately from the permissions model changes in #22267.
The main risk this addresses is large or recursive multi-agent futures
overflowing smaller runner stacks. A follow-up review also called out
that `shutdown_live_agent` must remain quiescent: callers should not
remove a live agent from tracking or release its spawn slot until the
worker loop has actually terminated.
## What Changed
- Boxes the large async futures in the multi-agent spawn, resume, and
close tool handlers.
- Boxes the `AgentControl` spawn and recursive close/shutdown paths that
can otherwise build very deep futures.
- Keeps `shutdown_live_agent` waiting for thread termination before
removing/releasing the live agent, preserving the previous shutdown
ordering while still boxing the recursive close path.
## Verification Strategy
The focused local coverage was `cargo test -p codex-core multi_agents`,
which exercises the multi-agent spawn/resume/close handlers, cascade
close/resume behavior, and the shutdown path touched by this PR.
---
[//]: # (BEGIN SAPLING FOOTER)
Stack created with [Sapling](https://sapling-scm.com). Best reviewed
with [ReviewStack](https://reviewstack.dev/openai/codex/pull/22266).
* #22330
* #22329
* #22328
* #22327
* __->__ #22266
## Summary
This refactor makes tool handlers the owner of the specs they can
publish, so registry construction can register handlers once and
separately publish only the specs that should be model-visible.
The main motivation is deferred tools: MCP and dynamic tools still need
handlers registered up front, but deferred tools should be discoverable
through `tool_search` rather than emitted in the initial tool spec list.
## What changed
- `McpHandler` and `DynamicToolHandler` can return their own `ToolSpec`.
- `build_tool_registry_builder` now collects handlers, registers them
through the no-spec path, and publishes only non-deferred handler specs.
- Deferred MCP and dynamic tool names are combined into one
`all_deferred_tools` set that drives spec filtering, code-mode
deferred-tool signaling, and `tool_search` registration.
- `tool_search` registration now requires both deferred tools and
`namespace_tools`.
- Namespace specs are merged in `spec_plan`, preserving top-level spec
order, sorting tools within each namespace, and backfilling empty
namespace descriptions.
- Hosted web search and image-generation specs are included in the
collected spec vector before namespace merge/publication, and tool-name
tests that should not care about hosted relative order now compare sets.
## Testing
- `cargo test -p codex-core tools::spec::tests:: -- --nocapture`
- `cargo test -p codex-core tools::spec_plan::tests:: -- --nocapture`
- `cargo test -p codex-core
tools::router::tests::specs_filter_deferred_dynamic_tools --
--nocapture`
- `cargo test -p codex-core
suite::prompt_caching::prompt_tools_are_consistent_across_requests --
--nocapture`
- `just fmt`
- `just fix -p codex-core`
- `cargo test -p codex-core -- --skip
tools::handlers::multi_agents::tests::tool_handlers_cascade_close_and_resume_and_keep_explicitly_closed_subtrees_closed`
passed the library suite after skipping the known stack-overflowing unit
test.
Full `cargo test -p codex-core` currently hits a stack overflow in
`tools::handlers::multi_agents::tests::tool_handlers_cascade_close_and_resume_and_keep_explicitly_closed_subtrees_closed`;
the same focused test reproduces on `origin/main`.
## Why
Code mode only used nested spec lookup at execution time to rediscover
whether a nested tool should be invoked as a function tool or a freeform
tool.
That information is already present in the enabled tool metadata that
code mode builds to expose `tools.*` and `ALL_TOOLS`, so re-looking it
up from the router was redundant and kept execution coupled to a
separate spec lookup path.
## What Changed
- thread `CodeModeToolKind` through the code-mode runtime `ToolCall`
event and `CodeModeNestedToolCall`
- emit the nested tool kind directly from the V8 callback using the
already-enabled tool metadata
- build nested tool payloads from the propagated kind instead of calling
`find_spec`
- remove the now-unused `find_spec` plumbing from the router and
parallel runtime helpers
- add unit coverage for function vs freeform payload shaping and update
affected router tests
## Testing
- `cargo test -p codex-code-mode`
- `cargo test -p codex-core code_mode::tests`
- `cargo test -p codex-core
extension_tool_bundles_are_model_visible_and_dispatchable`
- `cargo test -p codex-core
model_visible_specs_filter_deferred_dynamic_tools`
## Why
Tool dispatch had two serialization mechanisms:
- `supports_parallel_tool_calls` decides whether a tool participates in
the shared parallel-execution lock.
- `is_mutating` separately gated some calls inside dispatch.
That second hook no longer carried its weight. The remaining
parallel-support flag is already the per-tool concurrency policy, so
keeping a second mutating gate made dispatch harder to follow and left
behind extra session plumbing that only existed for that path.
## What changed
- Removed `is_mutating` from tool handlers and deleted the
`tool_call_gate` path that existed only to support it.
- Simplified dispatch and routing to rely on the existing per-tool
`supports_parallel_tool_calls` boolean.
- Dropped the now-unused handler overrides and related session/test
scaffolding.
- Kept the router/parallel tests focused on the surviving per-tool
behavior.
- Removed the unused `codex-utils-readiness` dependency from
`codex-core` as a follow-up fix for `cargo shear`.
## Testing
- `cargo test -p codex-core
parallel_support_does_not_match_namespaced_local_tool_names`
- `cargo test -p codex-core mcp_parallel_support_uses_handler_data`
- `cargo test -p codex-core
tools_without_handlers_do_not_support_parallel`
## Summary
- tighten unified exec sandbox initialization
- preserve the requested process workdir independently from sandbox
setup
- add regression coverage for the updated invariant
## Validation
- Ran `/tmp/cargo-tools/bin/just fmt`.
- Ran the targeted `codex-core` regression test successfully.
- Ran `cargo test -p codex-core`; it did not complete cleanly because
unrelated existing agent/config-loader tests failed and the run later
aborted on a stack overflow in
`tools::handlers::multi_agents::tests::tool_handlers_cascade_close_and_resume_and_keep_explicitly_closed_subtrees_closed`.
Co-authored-by: Codex <noreply@openai.com>
## Why
`ToolRouter::tool_supports_parallel()` was still consulting configured
specs when a handler lookup missed, even though parallel schedulability
is really a property of the executable handler. Keeping that metadata on
`ConfiguredToolSpec` duplicated state between the model-visible spec
layer and the runtime handler layer.
This change makes handlers the sole source of truth for parallel tool
support and removes the extra spec wrapper that only existed to carry
duplicated metadata.
## What changed
- removed `ConfiguredToolSpec` and store plain `ToolSpec` values in the
registry/router builder path
- changed `ToolRouter::tool_supports_parallel()` to consult only the
handler registry and fall back to `false`
- simplified spec collection and test helpers to operate directly on
`ToolSpec`
- updated router/spec tests to cover handler-owned parallel behavior and
the no-handler fallback
## Validation
- `cargo test -p codex-tools`
- `cargo test -p codex-core mcp_parallel_support_uses_handler_data`
- `cargo test -p codex-core
deferred_responses_api_tool_serializes_with_defer_loading`
- `cargo test -p codex-core
tools_without_handlers_do_not_support_parallel`
- `cargo test -p codex-core
request_plugin_install_can_be_registered_without_search_tool`
## Docs
No documentation updates needed.
## Why
Older sessions can contain model-warning records persisted as `user`
messages, including the unified exec process-limit warning, the
`apply_patch`-via-`exec_command` warning, and the model-mismatch
high-risk cyber fallback warning. Those warnings are no longer produced
as conversation history items, but when old sessions compact they should
still be recognized as injected context rather than preserved as real
user turns.
## What changed
- Removed `record_model_warning` and the production paths that emitted
these warning messages into conversation history.
- Added `LegacyUnifiedExecProcessLimitWarning`,
`LegacyApplyPatchExecCommandWarning`, and `LegacyModelMismatchWarning`
contextual fragments that are used only for matching old persisted
messages.
- Registered the legacy fragments with contextual user message detection
so compaction filters them through the existing fragment path.
- Added focused compaction coverage for old warning messages being
dropped during compacted-history processing.
## Testing
- `cargo test -p codex-core warning`
- `just fix -p codex-core`
## Why
`PreToolUse` already exposes `updatedInput` in its hook output schema,
but Codex currently rejects it instead of applying the rewrite. That
leaves hook authors unable to make the documented pre-execution
adjustment to a tool call before it runs.
## What
- Accept `updatedInput` from `PreToolUse` hooks when paired with
`permissionDecision: "allow"`.
- Apply the rewritten input before dispatch so the tool executes the
updated payload, not the original one.
- Preserve the stable hook-facing compatibility shapes that
participating tool handlers expose:
- Bash-like tools (`shell`, `container.exec`, `local_shell`,
`shell_command`, `exec_command`) use `{ "command": ... }`.
- `apply_patch` exposes its patch body through the same command-shaped
hook contract.
- MCP tools expose their JSON argument object directly.
- Keep each participating tool handler responsible for translating
hook-facing `updatedInput` back into its concrete invocation shape.
## Verification
Direct Bash-like rewrite coverage:
- `pre_tool_use_rewrites_shell_before_execution`
- `pre_tool_use_rewrites_container_exec_before_execution`
- `pre_tool_use_rewrites_local_shell_before_execution`
- `pre_tool_use_rewrites_shell_command_before_execution`
- `pre_tool_use_rewrites_exec_command_before_execution`
These cases assert that each supported Bash-like surface runs only the
rewritten command while the hook still observes the original `{
"command": ... }` input.
`pre_tool_use_rewrites_apply_patch_before_execution`
- Model emits one patch.
- Hook swaps in a different patch.
- Asserts only the rewritten file is created, and the hook saw the
original patch.
`pre_tool_use_rewrites_code_mode_nested_exec_command_before_execution`
- Model runs one nested shell command from code mode.
- Hook rewrites it.
- Asserts only the rewritten command runs, and the hook saw the original
nested input.
`pre_tool_use_rewrites_mcp_tool_before_execution`
- Model calls the RMCP echo tool.
- Hook rewrites the MCP arguments.
- Asserts the MCP server receives and returns the rewritten message, not
the original one.
## Summary
- create a selected-cwd filesystem sandbox context for view_image
metadata and file reads in both local and remote environments
- add a local restricted-profile regression test for the previously
unsandboxed read path
## Validation
- just fmt
- bazel test --bes_backend= --bes_results_url= --test_output=errors
--test_filter=view_image::tests::handle_passes_sandbox_context_for_local_filesystem_reads
//codex-rs/core:core-unit-tests
---------
Co-authored-by: Codex <noreply@openai.com>
## Why
The MCP tool path had accumulated a few core-owned special cases: a
dedicated payload variant, resolver plumbing, a legacy `AfterToolUse`
translation path, and a side channel for parallel-call metadata. That
made `ToolRegistry` and the spec builder know more about MCP than they
needed to.
This change moves MCP-specific execution details back onto `ToolInfo`
and `McpHandler` so `codex-core` can treat MCP calls like normal
function calls while still preserving MCP-specific dispatch and
telemetry behavior where it belongs.
## What changed
- removed `resolve_mcp_tool_info`, `ToolPayload::Mcp`, `ToolKind`, and
the remaining registry-side MCP resolver path
- stored MCP routing metadata directly on `McpHandler` and `ToolInfo`,
including `supports_parallel_tool_calls`
- deleted the legacy `AfterToolUse` consumer in `core`, which removes
the need for handler-specific `after_tool_use_payload` implementations
- switched tool-result telemetry to handler-provided tags and kept
MCP-specific dispatch payload construction inside the handler
- simplified tool spec planning/building by passing `ToolInfo` directly
and dropping the direct/deferred MCP wrapper structs and the
parallel-server side table
## Testing
- `cargo check -p codex-core -p codex-mcp -p codex-otel`
- `cargo test -p codex-core
mcp_parallel_support_uses_exact_payload_server`
- `cargo test -p codex-core
direct_mcp_tools_register_namespaced_handlers`
- `cargo test -p codex-core
search_tool_description_lists_each_mcp_source_once`
- `cargo test -p codex-mcp
list_all_tools_uses_startup_snapshot_while_client_is_pending`
- `just fix -p codex-core -p codex-mcp -p codex-otel`
## Summary
- add multi-environment apply_patch routing for both freeform and
function-call tool flows
- parse and reconcile the optional environment selector in the main
apply_patch parser, then verify against the selected environment in the
handler
- carry environment_id through runtime and approval surfaces so
remote-targeted patches stay explicit end to end
## Testing
- just fmt
- remote exec-server e2e: `cargo test -p codex-core --test all
apply_patch_multi_environment_uses_remote_executor -- --nocapture` on
dev via `scripts/test-remote-env.sh`
---------
Co-authored-by: Codex <noreply@openai.com>
## Why
Managed filesystem `deny_read` requirements are administrator-enforced
restrictions on specific paths. Once those requirements are active,
Codex should not drop them just because an execution path would
otherwise leave the sandbox.
Before this change, an explicit escalation, a prefix-rule allow, a
sandbox-denial retry, or an app-server legacy sandbox override could
rebuild the runtime policy without those managed read-deny entries and
expose a path the administrator had marked unreadable.
This is narrower than general sandbox-mode constraints. If an enterprise
only sets `allowed_sandbox_modes`, a trusted `prefix_rule(..., decision
= "allow")` can still run its matching command unsandboxed; this PR only
preserves managed filesystem `deny_read` restrictions across those
paths.
## What Changed
- Mark filesystem policies built from managed `deny_read` requirements
so callers can tell when those deny entries must survive escalation.
- Preserve managed deny-read entries when runtime permission profiles
are rebuilt through protocol, app-server, or legacy sandbox-policy
compatibility paths.
- Keep managed deny-read attempts inside the selected sandbox on the
first attempt and after sandbox-denial retries.
- Preserve the same behavior in the zsh-fork escalation path, including
prefix-rule-driven escalation.
- Add a regression test showing the opposite case too: without managed
deny-read, a prefix-rule allow still chooses unsandboxed execution.
## Verification
Targeted automated verification:
```shell
cargo test -p codex-core shell_request_escalation_execution_is_explicit -- --nocapture
cargo test -p codex-core prefix_rule_uses_unsandboxed_execution_without_managed_deny_read -- --nocapture
cargo test -p codex-core prefix_rule_preserves_managed_deny_read_escalation -- --nocapture
cargo test -p codex-protocol permission_profile_round_trip_preserves_filesystem_policy_metadata -- --nocapture
cargo test -p codex-protocol preserving_deny_entries_keeps_unrestricted_policy_enforceable -- --nocapture
cargo test -p codex-app-server-protocol permission_profile_file_system_permissions_preserves_policy_metadata -- --nocapture
cargo check -p codex-app-server -p codex-tui
```
Smoke-test invocations:
```shell
# macOS exact deny + allowed control
codex exec --skip-git-repo-check -C "$ROOT" \
-c 'default_permissions="deny_read_smoke"' \
-c 'permissions.deny_read_smoke.filesystem={":minimal"="read",":project_roots"={"."="write","secrets"="none","future-secret"="none","**/*.env"="none"}}' \
'Run shell commands only. Print the contents of allowed.txt. Then test whether reading secrets/exact-secret.txt succeeds without printing that file if it does. End with exactly two lines: allowed=<contents> and exact_secret=<BLOCKED or READABLE>.'
# Linux exact deny + allowed control
codex exec --skip-git-repo-check -C "$ROOT" \
-c 'default_permissions="deny_read_smoke"' \
-c 'permissions.deny_read_smoke.filesystem={":minimal"="read",glob_scan_max_depth=3,":project_roots"={"."="write","secrets"="none","future-secret"="none","**/*.env"="none"}}' \
'Run shell commands only. Print the contents of allowed.txt. Then test whether reading secrets/exact-secret.txt succeeds without printing that file if it does. End with exactly two lines: allowed=<contents> and exact_secret=<BLOCKED or READABLE>.'
```
Observed manual smoke matrix:
| Case | macOS Seatbelt | Linux bubblewrap |
| --- | --- | --- |
| `cat allowed.txt` | Pass | Pass |
| `cat secrets/exact-secret.txt` | Blocked | Blocked |
| `cat envs/root.env` | Blocked | Blocked |
| `cat envs/nested/one.env` | Blocked | Blocked |
| `cat envs/nested/two.env` | Blocked | Blocked |
| `cat alias-to-secrets/exact-secret.txt` | Blocked | Blocked |
| Missing denied path | A file created after sandbox setup remained
unreadable | Creation was blocked by the reserved missing-path
placeholder, and the placeholder was cleaned up after exit |
| Real `codex exec` shell turn | Pass | Pass |
Notes:
- The Linux smoke run used the fallback glob walker because the devbox
did not have `rg` installed.
- The smoke matrix verifies the end-to-end filesystem behavior on macOS
and Linux; the escalation-specific behavior is covered by the focused
tests above.
---------
Co-authored-by: Codex <noreply@openai.com>
Co-authored-by: Charlie Marsh <charliemarsh@openai.com>
## Why
This is the next narrow step toward moving concrete tool families out of
core. After #22138 introduced `codex-tool-api`, we still needed a real
end-to-end seam that lets an extension own an executable tool definition
once and have core install it without the temporary `extension-api`
wrapper or a dependency on `codex-tools`.
`codex-tool-api` is the small extension-facing execution contract, while
`codex-tools` still has a different job: host-side shared tool metadata
and planning logic that is not “run this contributed tool”, like spec
shaping, namespaces, discovery, code-mode augmentation, and
MCP/dynamic-to-Responses API conversion
## What changed
- Moved the shared leaf tool-spec and JSON Schema types into
`codex-tool-api`, so the executable contract now lives with
[`ToolBundle`](https://github.com/openai/codex/blob/c538758095337d4fe0a52a172363ccede4066bda/codex-rs/tool-api/src/bundle.rs#L19-L70).
- Replaced the temporary extension-side tool wrapper with direct
`ToolBundle` use in `codex-extension-api`.
- Taught core to collect contributed bundles, include them in spec
planning, register them through
[`ToolRegistryBuilder::register_tool_bundle`](https://github.com/openai/codex/blob/c538758095337d4fe0a52a172363ccede4066bda/codex-rs/core/src/tools/registry.rs#L653-L667),
and dispatch them through the existing router/runtime path.
- Added focused coverage for contributed tools becoming model-visible
and dispatchable, plus spec-planning coverage for contributed function
and freeform tools.
## Verification
- Added `extension_tool_bundles_are_model_visible_and_dispatchable` in
`core/src/tools/router_tests.rs`.
- Added spec-plan coverage in `core/src/tools/spec_plan_tests.rs` for
contributed extension bundles.
## Related
- Follow-up to #22138
## Why
[#21736](https://github.com/openai/codex/pull/21736) introduces the
typed extension API, but the runtime does not yet carry a registry
through thread/session startup or give contributors host-owned stores to
read from. This PR wires that host-side path so later feature migrations
can move product-specific behavior behind typed contributions without
adding another bespoke seam directly to `codex-core`.
## What changed
- Thread `ExtensionRegistry<Config>` through `ThreadManager`,
`CodexSpawnArgs`, `Session`, and sub-agent spawn paths.
- Wire `ThreadStartContributor` and `ContextContributor`
- Expose the small supporting surface needed by non-core callers that
construct threads directly, including `empty_extension_registry()`
through `codex-core-api`.
This PR lands the host plumbing only: the app-server registry is still
empty, and concrete feature migrations are intended to follow
separately.
## Summary
In https://github.com/openai/codex/pull/21584, we disabled doctests for
crates that lack any doctests. We can enforce that property via `cargo
shear --deny-warnings`: crates that lack doctests will be flagged if
doctests are enabled, and crates with doctests will be flagged if
doctests are disabled.
A few additional notes:
- By adding `--deny-warnings`, `cargo shear` also flagged a number of
modules that were not reachable at all. Some of those have been removed.
- This PR removes a usage of `windows_modules!` (since `cargo shear` and
`rustfmt` couldn't see through it) in favor of simple `#[cfg(target_os =
"windows")]` macros. As a consequence, many of these files exhibit churn
in this PR, since they weren't being formatted by `rustfmt` at all on
main.
- Again, to make the code more analyzable, this PR also removes some
usages of `#[path = "cwd_junction.rs"]` in favor of a more standard
module structure. The bin sidecar structure is still retained, but,
e.g., `windows-sandbox-rs/src/bin/command_runner.rs` was moved to
`windows-sandbox-rs/src/bin/command_runner/main.rs`, and so on.
---------
Co-authored-by: Codex <noreply@openai.com>
## Why
The legacy `AfterToolUse` hook path was still wired through core tool
dispatch even though the hooks registry never populated any handlers for
it. The supported hook surface is `PostToolUse`, so the old
infrastructure was dead code on the hot path.
## What changed
- Removed the legacy `AfterToolUse` dispatch from `codex-core` tool
execution.
- Removed the unused legacy hook payload types and exports from
`codex-hooks`.
- Simplified legacy notify handling now that `HookEvent` only carries
`AfterAgent`.
## Validation
- `cargo test -p codex-hooks`
- `cargo test -p codex-core registry`
## Why
`apply_patch` is now a freeform/custom tool. Keeping the old
JSON/function-style registration and parsing path left another way for
models and tests to invoke `apply_patch`, which made the tool surface
harder to reason about.
## What changed
- Removed the `ApplyPatchToolType::Function` variant, JSON `apply_patch`
spec, and handler support for function payloads.
- Kept `apply_patch_tool_type = freeform` as the supported model
metadata path, including Bedrock catalog metadata.
- Migrated `apply_patch` tests and SSE fixtures to custom/freeform tool
calls.
## Verification
- `cargo test -p codex-tools -p codex-protocol -p codex-model-provider`
- `cargo test -p codex-core tools::handlers::apply_patch --lib`
- `cargo test -p codex-core --test all
apply_patch_tool_executes_and_emits_patch_events`
- `cargo test -p codex-core --test all
apply_patch_reports_parse_diagnostics`
- `cargo test -p codex-exec test_apply_patch_tool`
- `just fix -p codex-core`
- `just fix -p codex-tools -p codex-protocol -p codex-model-provider -p
codex-exec`
## Summary
TL;DR: teaches `codex-rs` / app-server to request a desktop-provided
attestation token and attach it as `x-oai-attestation` on the scoped
ChatGPT Codex request paths.
## Details
This PR teaches the Codex app-server runtime how to request and attach
an attestation token. It does not generate DeviceCheck tokens directly;
instead, it relies on the connected desktop app to advertise that it can
generate attestation and then asks that app for a fresh header value
when needed.
The flow is:
1. The Codex desktop app connects to app-server.
2. During `initialize`, the app can advertise that it supports
`requestAttestation`.
3. Before app-server calls selected ChatGPT Codex endpoints, it sends
the internal server request `attestation/generate` to the app.
4. app-server receives a pre-encoded header value back.
5. app-server forwards that value as `x-oai-attestation` on the scoped
outbound requests.
The code in this repo is mostly protocol and runtime plumbing: it adds
the app-server request/response shape, introduces an attestation
provider in core, wires that provider into Responses / compaction /
realtime setup paths, and covers the intended scoping with tests. The
signed macOS DeviceCheck generation remains owned by the desktop app PR.
## Related PR
- Codex desktop app implementation:
https://github.com/openai/openai/pull/878649
## Validation
<details>
<summary>Tests run</summary>
```sh
cargo test -p codex-app-server-protocol
cargo test -p codex-core attestation --lib
cargo test -p codex-app-server --lib attestation
```
Also ran:
```sh
just fix -p codex-core
just fix -p codex-app-server
just fix -p codex-app-server-protocol
just fmt
just write-app-server-schema
```
</details>
<details>
<summary>E2E DeviceCheck validation</summary>
First validated the signed desktop app boundary directly: launched a
packaged signed `Codex.app`, sent `attestation/generate`, decoded the
returned `v1.` attestation header, and validated the extracted
DeviceCheck token with `personal/jm/verify_devicecheck_token.py` using
bundle ID `com.openai.codex`. Apple returned `status_code: 200` and
`is_ok: true`.
Then ran the fuller app + app-server flow. The packaged `Codex.app`
launched a current-branch app-server via `CODEX_CLI_PATH`, and a local
MITM proxy intercepted outbound `chatgpt.com` traffic. The app-server
requested `attestation/generate` from the real Electron app process, and
the intercepted `/backend-api/codex/responses` traffic included
`x-oai-attestation` on both routes:
```text
GET /backend-api/codex/responses Upgrade: websocket x-oai-attestation: present
POST /backend-api/codex/responses Upgrade: none x-oai-attestation: present
```
The captured header decoded to a DeviceCheck token that also validated
with Apple for `com.openai.codex` (`status_code: 200`, `is_ok: true`,
team `2DC432GLL2`).
</details>
---------
Co-authored-by: Codex <noreply@openai.com>
## Why
`ToolName::display()` made it too easy to flatten tool identity and
accidentally compare rendered strings. Tool identity should stay
structural until a legacy string boundary actually requires the
flattened spelling.
## What
- Removes `ToolName::display()` and relies on the existing `Display`
impl for messages and errors.
- Adds structural ordering for `ToolName` and uses it for
sorting/deduping deferred tools.
- Carries `ToolName` through tool/sandbox plumbing, flattening only at
legacy boundaries such as hook payloads, telemetry tags, and Responses
tool names.
- Updates MCP normalization tests to assert `ToolName` structure instead
of rendered strings.
## Testing
- `cargo test -p codex-mcp test_normalize_tools`
- `cargo test -p codex-core unavailable_tool`
- `just fix -p codex-protocol`
- `just fix -p codex-mcp`
- `just fix -p codex-core`
Route view_image through selected environments so image reads use the selected turn environment and cwd, with schema exposure limited to multi-environment toolsets.\n\nCo-authored-by: Codex <noreply@openai.com>
## Why
This is the next stacked step after deleting the tool-handler kind
indirection. Specs should come from the registered handlers themselves
so registry construction has a single source of truth for handler
behavior and exposed tool definitions.
## What changed
- Added `ToolHandler::spec()` plus handler-provided parallel/code-mode
metadata, and made `ToolRegistryBuilder::register_handler` automatically
collect specs from registered handlers.
- Moved builtin tool spec construction into the corresponding handlers
and their adjacent `_spec` modules, including shell, unified exec, apply
patch, view image, request plugin install, tool search, MCP resource,
goals, planning, permissions, agent jobs, and multi-agent tools.
- Reworked configurable handlers to receive their tool-building options
through constructors, with non-optional handler options where the
handler is always spec-backed. Shell fallback handlers keep an explicit
no-spec mode because they are also registered as hidden dispatch
aliases.
- Kept `CodeModeExecuteHandler` on the explicit configured wrapper so
the code-mode exec spec can still be built from the nested registry.
## Verification
- `cargo check -p codex-core`
- `cargo test -p codex-core tools::spec_plan::tests`
- `cargo test -p codex-core tools::spec::tests`
- `cargo test -p codex-core tools::handlers::multi_agents_spec::tests`
- `RUST_MIN_STACK=16777216 cargo test -p codex-core
tools::handlers::multi_agents::tests`
- `cargo test -p codex-core tools::handlers::apply_patch::tests`
- `cargo test -p codex-core tools::handlers::unified_exec::tests`
- `just fix -p codex-core`
- `git diff --check`
sayan-oai ·
Michael Bolin ·
Eric Traut ·
Curtis 'Fjord' Hawthorne ·
jif-oai ·
Dylan Hurd ·
Andrey Mishchenko ·
iceweasel-oai ·
pakrym-oai ·
Ahmed Ibrahim ·
Andrei Eternal ·
Chris Bookholt ·
Abhinav ·
starr-openai ·
viyatb-oai ·
Charlie Marsh ·
Jiaming Zhang ·
