## What changed
- Retain bounded, non-completion agent messages in remote v2 compaction history so delegated tasks remain available on follow-up turns.
- Account for encrypted agent-message content when estimating token usage and place restored initial context before the latest real user or agent message.
- Exclude child completion messages from retained follow-up history and strip inherited parent agent messages when forking a child.
## Testing
- Extend compaction, history-estimation, and agent-fork tests to cover encrypted delegated tasks, child completions, context ordering, and fork sanitization.
GitOrigin-RevId: ba4392e9cfdf4ecbcdc9c7d8581d58a8d2d78260
## What changed
- Preserve `encrypted_function_args` on function calls so an empty list can mark plaintext collaboration arguments across request replay.
- Deliver `spawn_agent`, `send_message`, and `followup_task` payloads as structured plaintext agent messages when that marker is present; retain encrypted delivery otherwise.
- Redact plaintext collaboration arguments from tool and communication logs, and omit the metadata when sending requests to non-OpenAI providers.
## Testing
- Cover serialization of empty encrypted-argument metadata and plaintext versus encrypted subagent message delivery.
- Verify plaintext tool arguments are redacted and provider-specific metadata is removed from non-OpenAI requests.
GitOrigin-RevId: 64db98ff0b61a3af2f04ed609292363f2e2362a8
## What changed
- Add protocol types and `ResponseItem` helpers for attaching and clearing
locally recorded tool-call names and arguments.
- Bound serialized metadata across a prompt, replacing oversized arguments and
omitted calls with trusted truncation details.
- Prevent deserialized response items or model-provided arguments from forging
locally generated tool-call and truncation metadata.
## Testing
- Cover prompt-wide size limits, omission accounting, idempotent bounding, and
forged truncation markers.
GitOrigin-RevId: 960a223437ce81fe6b4cbb5fc7a9774408b2755b
## What changed
- Add model and personality sections to the persisted world-state snapshot.
- Generate model-switch and personality instructions through world-state diffs, including when prior state must be inferred during replay.
- Keep those instruction fragments in standalone developer messages instead of merging them with adjacent context updates.
## Testing
- Cover initial, unchanged, and changed model and personality states.
- Verify that model changes are recorded in rollout world-state items.
GitOrigin-RevId: 497e29cf38cac08e992e546df900f6f8b01120f9
## Why
Multi-agent mode instructions are durable model context and need to survive
history changes without re-emitting unrelated setup hints.
## What changed
- Add a `multi_agent_mode` world-state section that diffs, retains, and restores
the effective mode instructions.
- Persist mode changes in world-state snapshots while keeping the legacy
`TurnContextItem.multi_agent_mode` field for reading older rollouts.
- Keep initial mode instructions after the root-agent usage hint so the active
mode takes precedence, and bound custom mode text before snapshotting it.
## Testing
- Add snapshot coverage for mode transitions, retained history, and custom-text
truncation.
- Verify live mode changes preserve instruction ordering, avoid duplicating the
usage hint, and record both modes in world state.
GitOrigin-RevId: 96b9964d6e5299b07a011c559b68e3d948a49226
## Why
Remote compaction can process large histories. Repeatedly estimating and replacing the full history, and cloning it when tracing is disabled, adds avoidable CPU and memory overhead.
## What changed
- Estimate each history item's token count once, update the total as trailing tool outputs are rewritten, and replace history only after all rewrites are selected.
- Preserve unclamped token totals while calculating removed tokens so saturated estimates do not hide overflow.
- Snapshot compaction input history only when rollout tracing is enabled.
- Reuse the v2 request input instead of cloning it before adding the compaction trigger.
## Testing
- Cover enabled and disabled compaction trace contexts through `is_enabled()` assertions.
GitOrigin-RevId: 8de9c9704ba29532fba05430eb204bbae2e8bf83
## What changed
- Assert truncated audio history through `ContextManager::raw_items()`.
- Initialize `additional_context_limit` in the quoted Windows hook command test.
GitOrigin-RevId: 43c19804e558b7a11fc17590506ec6ec7f3a9897
## Why
Cloning a `ContextManager` deep-copies every `ResponseItem`, even when callers only need a read-only history snapshot.
## What changed
- Store history items in an `Arc<Vec<ResponseItem>>` so cloned snapshots share their items until one is mutated.
- Reuse an owned vector when consuming an unshared snapshot and clone it only when other snapshots still reference it.
- Estimate session token counts from a snapshot after releasing the session-state lock.
## Testing
Added a regression test that verifies cloned histories share their backing items before mutation and diverge without affecting each other afterward.
GitOrigin-RevId: 24605888209ae9ff78cadb04d9c36463f7e2dbe0
## Why
Audio attachments could be dropped from tool outputs and reconstructed thread history, while inline base64 payload size could overstate their model context cost.
## What changed
- Preserve remote and local audio attachments in legacy user-message events and thread history.
- Keep audio from function and custom tool outputs when the model supports it and the output fits the truncation budget; otherwise replace it with an omission marker.
- Estimate inline audio cost from its duration, with a serialized-size fallback for malformed or unsupported data URLs, so compaction and truncation use model-visible cost instead of raw base64 size.
- Replace audio in messages and tool outputs with an explanatory placeholder for models without audio input support.
## Testing
Add coverage for history round trips, dynamic tool responses, modality filtering, duration-based context estimates, output-budget truncation, and remote compaction.
GitOrigin-RevId: 5ed4628850a668f8383b7cd73db71b144cc02cd1
## What changed
Remove the fallback that replaced image content in the latest tool output with
`Invalid image` and retried the model request. Invalid-image responses now use
the existing bad-request error path without modifying conversation history.
GitOrigin-RevId: f4ea782f16c5451feb261694caa73b77f24c3e4e
## What changed
- Tell the model that generated images are already displayed to the user and should not be repeated as Markdown images or file links.
- Keep generated-image artifact paths and output hints owned by the image generation extension.
- Centralize the availability check for the image generation tool across feature, provider, namespace-tool, model modality, and authentication requirements.
## Testing
- Verify artifact paths sanitize session and call identifiers.
- Verify the model-facing output hint reports that the generated image is already displayed.
GitOrigin-RevId: 6c865b6a96d6a71bb2c33246d1518a00e7c8b840
## What changed
- Add `audio` to the model input modalities exposed by the protocol and generated app-server schemas.
- Preserve audio in prompts for models that advertise audio input, and replace historical audio with an omission marker for models that do not.
## Testing
- Cover prompt normalization with and without audio support.
- Extend the model-switching integration test to verify that switching from a multimodal model to a text-only model strips prior image and audio content.
GitOrigin-RevId: c18e47ad602720216303ff9abb5bca4d508391ee
## What changed
- Model permission instructions as a world-state section keyed by a stable hash of their rendered developer message.
- Re-emit permission context when its model-visible contents change or the retained fragment is missing, while avoiding duplicates when matching instructions already exist in history, including bundled developer messages.
- Normalize CRLF line endings when hashing world-state fragments so equivalent content produces the same snapshot.
## Testing
- Add snapshot coverage for unchanged, changed, absent, and unknown permission state.
- Cover bundled retained permissions and restored world-state baselines during rollout reconstruction.
GitOrigin-RevId: 7d8d212460c1690ca617859f66dcd825987f1e23
## What changed
- Represent whether realtime is active as a persisted world-state section.
- Emit start and end instructions from world-state transitions, including when
reconstructing context after resume.
- Retain existing realtime start fragments and avoid emitting another start when
custom instructions change while realtime remains active.
## Testing
- Add snapshot coverage for absent, unknown, active, and inactive realtime state
transitions.
- Cover realtime start and end updates plus resuming with changed custom start
instructions.
GitOrigin-RevId: ec670b8132b0f1c9317738bfdeb272069927d91d
## What changed
- Model collaboration mode as a world-state section whose persisted snapshot is
the active mode.
- Emit collaboration instructions when the mode changes, while ignoring
instruction text changes within the same mode.
- Restore persisted collaboration instructions when they are missing from
retained history, including after a fork.
## Testing
- Add snapshot coverage for initial, unchanged, changed, and unknown mode state.
- Cover retained-history restoration and same-mode instruction changes.
GitOrigin-RevId: ffd1683499f09acef6516b0bfcb5bcb4d3f28df4
## What changed
- Treat an empty custom `multi_agent_mode_hint_text` as having no instructions to emit.
- Avoid adding an empty multi-agent mode message both on the initial turn and when the configured hint changes after resuming a session.
## Testing
- Update the empty-hint integration test to assert that no mode message is sent.
- Add resume coverage that verifies changing a custom hint to an empty string emits no update.
GitOrigin-RevId: 8033292115936fac2cf22f85141278595804ee61
## Why
An execution environment marked `starting` is not yet usable, but unrelated work can continue while it becomes ready.
## What changed
- Add developer guidance that explains execution-environment isolation and the resources that become available after startup.
- Tell the model to wait only when the current task needs the starting environment.
- Retain the guidance across world-state updates without duplicating it in request history.
## Testing
- Cover world-state transitions, history retention, rollback handling, and deferred-environment request deduplication.
GitOrigin-RevId: f12f9204d4cbddce50a70eb0998beb6b66cd2fa9
## What changed
- Add per-sandbox-mode permission messages to `ModelMessages` and preserve them when applying model configuration overrides.
- Use the selected catalog message in permission instructions, substituting `{{ network_access }}` with the active network policy. Fall back to the existing sandbox text when no override is provided, and allow an empty override to omit only the sandbox section.
- Apply catalog permission messages when a session starts and when its model changes.
## Testing
- Cover catalog deserialization, mode selection, network substitution, fallback and empty-message behavior, remote model catalogs, and model changes.
GitOrigin-RevId: cd5ed3aee3155dca0e7b2358c0f09bae73236856
## Why
`TurnContext` stored model and reasoning settings both as resolved turn fields and inside `CollaborationMode`, requiring the copies to stay synchronized when the model changed.
## What changed
- Store the mode and collaboration-mode developer instructions directly on `TurnContext`.
- Reconstruct `CollaborationMode` from those values plus the turn's current model and reasoning effort when callers need it.
- Read the mode directly for plan-mode checks, lifecycle events, and telemetry.
GitOrigin-RevId: b8fccafdd5d5cbccc84bc70d542588087ddf0093
## What changed
- Add a `ResponseItemId` type that generates item-specific prefixes with UUIDv7 suffixes and use it across response items.
- Keep deserialization permissive for legacy histories, but omit empty or unprefixed item IDs from HTTP and WebSocket requests.
- Export the new type in the generated TypeScript protocol schema.
## Testing
- Cover prefixed ID generation, legacy deserialization, prefix recognition, and outbound request filtering.
GitOrigin-RevId: 0209fe430c826d0ae88bc4648652c4ebf390c4a4
## Why
A thread resumed without an explicit reviewer could pick up the reviewer
from the current config instead of preserving the reviewer already in
use by the thread. After an app restart, this meant a thread running
with auto review could silently switch back to user review, and the next
turn could continue under the wrong reviewer.
## What changed
Persist the effective reviewer with each turn and restore the latest
persisted value when the thread resumes. If the resume request
explicitly provides a reviewer, that value still takes precedence.
## Test plan
- Added a regression test that starts a thread with auto review, records
a turn, restarts with user review in config, resumes without an
override, and verifies that auto review is preserved.
- `just test -p codex-protocol`
- `just test -p codex-state`
- `just test -p codex-rollout`
- `just test -p codex-app-server
thread_resume_preserves_persisted_approvals_reviewer`
- Clippy for the affected crates
## Why
Approval guidance is currently assembled entirely by the client. Model
Messages V2 needs model catalogs to provide model-specific `on_request`
guidance for both user-reviewed and auto-reviewed approval flows while
retaining the existing generated prompt as a compatibility fallback.
## What changed
- add nullable `on_request` and `on_request_auto_review` catalog
messages
- select the message matching the active approvals reviewer for
`on_request` policies
- replace the complete legacy approval section when the selected catalog
value exists, including support for an empty string that suppresses the
section
- retain legacy rendering when the object or selected key is absent, and
for non-`on_request` policies
- preserve approval messages when base-instruction or personality
overrides clear instruction templates
- refresh permissions instructions when the active model changes
- pass catalog messages through initial and incremental permissions
construction
## Relationship to reviewer persistence
PR #31309 independently persists the approvals reviewer in turn context
and refreshes permissions when that reviewer changes. This PR is based
directly on `main` and does not duplicate that rollout migration; once
both land, reviewer switches will also select and append the new catalog
variant.
## Testing
- `just test -p codex-protocol`
- `just test -p codex-prompts`
- `just test -p codex-models-manager`
- `just test -p codex-core permissions_messages`
## Why
Generic Apps guidance is emitted only while building static initial
context. If the Apps MCP is unavailable then and recovers later in the
same turn, its tools can become usable without the model receiving the
guidance for using them.
## What
- move generic Apps guidance into a persisted `apps_instructions` World
State section
- derive availability from the request's MCP runtime while preserving
the existing feature, auth, orchestrator-MCP, and config gates
- recognize legacy and retained Apps fragments so resume and compaction
do not duplicate guidance
- register Apps guidance as rollback-trimmable context
- remove the old static injection path and its now-unused connector
helper
- keep tool construction on its existing independent `list_all_tools()`
read rather than adding a request-wide cache; a reconnect between the
two reads can differ for one request and reconciles on the next request
Apps and plugin guidance now render after the remaining static
host-skills block, with Apps before Plugins.
## Testing
- `just fmt`
- `just test -p codex-core apps_instructions`
- `just test -p codex-core
apps_guidance_appears_after_background_recovery_within_a_turn`
- `just test -p codex-core
drop_last_n_user_turns_trims_context_updates_above_rolled_back_turn`
## Why
Generic plugin guidance is currently emitted only with initial context
from host plugin state. An executor-selected plugin can become available
later in the same turn, making its skills and tools usable without ever
telling the model how plugin capabilities should be used.
## What
- project every ready selected plugin package, including skill-only
plugins
- carry plugin availability with the exact MCP runtime projection while
preserving MCP manager reuse when servers and connectors are unchanged
- move generic plugin guidance from the static initial-context path into
persisted World State
- recognize legacy and retained plugin fragments so resume and
compaction do not duplicate guidance
## Testing
- `just test -p codex-mcp-extension`
- `just test -p codex-core plugins_instructions`
- `just test -p codex-core
plugin_availability_change_reuses_the_mcp_manager`
- `just test -p codex-app-server --test all selected_capabilit`
## Why
Multi-agent V2 normally derives its mode instructions from reasoning
effort: Ultra enables proactive delegation, while other efforts require
an explicit request. Some deployments need to provide one configured
delegation policy that replaces those built-ins and remains stable when
reasoning effort changes.
## What changed
- Add `features.multi_agent_v2.multi_agent_mode_hint_text` alongside the
existing root and subagent hint settings.
- Treat any configured value, including an empty string, as
`MultiAgentMode::Custom(hint_text)`, so the configured text replaces the
built-in explicit-only and proactive policies.
- Persist the full custom variant and hint text in the turn-context
snapshot, so the durable comparison baseline detects both
reasoning-effort changes and configured policy-text changes.
- Preserve the existing explicit-only/proactive behavior when the
setting is absent.
- Replace the ambiguous `MultiAgentMode::None` variant with
`MultiAgentMode::Custom(String)` in new rollouts and API schemas. A
compatibility wire type maps legacy serialized `none` values to
`Custom("")` when resuming existing rollouts.
- Regenerate the config and app-server schemas.
## Configuration examples
The distinction is whether `multi_agent_mode_hint_text` is present. An
empty string is still a configured value and intentionally suppresses
the built-in mode instructions.
### Unset: preserve existing effort-derived behavior
```toml
[features.multi_agent_v2]
enabled = true
# multi_agent_mode_hint_text is omitted
```
- Ultra reasoning uses the built-in proactive delegation instructions.
- Other reasoning efforts use the built-in explicit-request-only
instructions.
### Empty: suppress all mode hint text
```toml
[features.multi_agent_v2]
enabled = true
multi_agent_mode_hint_text = ""
```
This selects effective mode `custom` at every reasoning effort and
injects an empty mode body, suppressing both built-in policies.
### Set: always use the configured text
```toml
[features.multi_agent_v2]
enabled = true
multi_agent_mode_hint_text = "Delegate to subagents when it will materially improve the result."
```
This selects effective mode `custom` at every reasoning effort and
injects the configured text verbatim instead of either built-in policy.
## Verification
- `just test -p codex-core multi_agent_mode`
- Covers a configured hint across High and Ultra reasoning efforts and
verifies the full custom hint is recorded for both turns.
- Covers an empty-string override suppressing both built-in instruction
bodies.
- `just test -p codex-protocol -p codex-app-server-protocol`
- Covers legacy `none` turn-context deserialization as `Custom("")` and
verifies the regenerated schemas.
## Why
Response item IDs represent stable conversation identity.
`ContextManager::for_prompt` repairs an unmatched call by synthesizing
an `"aborted"` output in the disposable prompt projection, but that
output previously had no ID. Assigning a fresh ID on every prompt build
would make retries and resumes change otherwise identical model context
and reduce prompt-cache reuse.
The concrete bug is that these normalization-created outputs bypass the
regular item-ID allocation path. Even with item IDs enabled, a prompt
could therefore contain an identified call paired with a synthetic
output whose `id` was missing. This change closes that gap by deriving
the output ID from the source call's item ID. For legacy calls that have
no item ID, the output remains ID-less because there is no stable source
identity to derive from.
The originating call already has a stable item ID under the item-ID
model introduced in #28814. A prompt-only output can therefore derive
stable identity from that call without mutating canonical history or
persisted rollouts. This addresses the failure exposed by #30311 while
keeping normalization read-only outside its detached prompt snapshot.
UUIDv5 is intentional here because it is the standard namespaced,
deterministic UUID construction. Using the output kind and source call
ID as the name produces the same UUID on every projection while keeping
output kinds in separate name domains. UUIDv7 would introduce randomness
and time, so keeping it stable would require persisting the synthetic
repair. UUIDv5 uses SHA-1 internally, but this is only an identity
mapping—not an authenticity or security boundary.
## What changed
- Derive a deterministic UUIDv5 ID for each synthesized call output from
the source call item ID.
- Use the Responses API prefix appropriate for function, custom-tool,
tool-search, and local-shell outputs.
- Preserve the existing insertion position immediately after the
unmatched call.
- Keep synthesized outputs prompt-only; no rollout, task-lifecycle,
compaction, or raw-response behavior changes.
## Testing
- `just test -p codex-core
for_prompt_assigns_stable_id_to_synthetic_output_without_reordering_history`
- `just test -p codex-core
synthetic_call_output_id_is_stable_across_resumes`
- `just test -p codex-core normalize_adds_missing_output`
- `just test -p codex-core response_item_ids`
## Summary
- Preserve the optional namespace on custom tool calls during response
deserialization and app-server replay.
- Use the namespaced tool identifier for streaming argument handling and
tool dispatch.
- Regenerate app-server protocol schemas.
- Add regression tests covering namespace serialization and routing.
## Testing
- Ran affected protocol and app-server test suites.
- Ran the full core test suite; two load-sensitive timing tests passed
when rerun individually.
- Ran Clippy and formatting checks.
- Verified with a local end-to-end app-server replay that the namespace
is preserved through the complete request/response flow.
## Why
Older rollouts can retain model-visible context for a WorldState section
without having a persisted snapshot for that section. Treating the
missing snapshot as definitely absent can duplicate old context or fail
to tell the model that it was replaced or removed.
This provides a generic migration path for sections moving into
WorldState, beginning with AGENTS.md.
Builds on #29810.
## What changed
- distinguish section state that is absent, known from a persisted
snapshot, or unknown because matching legacy context remains in history
- let WorldState sections identify their own legacy fragments while
`ContextManager` owns history reconciliation and baseline persistence
- make AGENTS.md emit one conservative replacement or removal update for
legacy history, then deduplicate from the newly persisted baseline
- preserve existing environment rendering when persisted section data is
missing or malformed
## Testing
- `just test -p codex-core world_state`
- `just test -p codex-core
cold_resume_invalidates_deleted_legacy_agents_md_once -- --exact`
## Why
Persisting `WorldState` snapshots and patches is only useful if resume
and fork restore that exact comparison baseline. Rebuilding it from
`TurnContextItem` loses section state and can either repeat or suppress
model-visible updates.
This is the third PR in the WorldState persistence stack, built on
#29835.
## What
- Replay full WorldState snapshots and RFC 7386 patches through the
existing rollout reconstruction segments.
- Discard state from rolled-back turns and treat compaction as a
baseline reset.
- Hydrate `ContextManager` from the reconstructed snapshot on resume and
fork.
- Remove the synthetic `TurnContextItem` to WorldState conversion path.
- Leave legacy or malformed rollouts without a baseline so the next
update safely emits a full snapshot.
## Testing
- `just test -p codex-core world_state`
- `just test -p codex-core rollout_reconstruction_tests`
- `just fix -p codex-core`
- `just test -p codex-core` *(the changed tests passed; the full run
also hit unrelated existing/test-environment failures, primarily a
missing `test_stdio_server` binary)*
## Why
Ultra should be one user-facing reasoning selection for work that
benefits from both maximum reasoning and proactive multi-agent
delegation. Without it, clients must coordinate maximum reasoning with
the experimental `multiAgentMode` setting, even though the inference
backend still expects its existing `max` effort value.
This change makes reasoning effort the source of truth: clients select
`ultra`, core derives proactive multi-agent behavior when the turn is
eligible for multi-agent V2, and inference requests continue to use the
backend-compatible `max` value.
## What changed
- Add `ultra` as a first-class reasoning effort and preserve
model-catalog ordering when exposing it to clients.
- Convert `ultra` to `max` at the inference request boundary, including
Responses HTTP/WebSocket requests, startup prewarm, compaction, and
memory summarization.
- Derive effective multi-agent mode per turn from effective reasoning
effort:
- eligible multi-agent V2 + `ultra` → `proactive`
- eligible multi-agent V2 + any other effort → `explicitRequestOnly`
- V1 or otherwise ineligible sessions → no multi-agent mode instruction
- Keep the derived effective mode in turn context history so successive
turns can emit a developer-message update only when the effective mode
changes.
- Remove selected multi-agent mode from core session configuration, turn
construction, thread settings, resume/fork restoration, and subagent
spawn plumbing. Subagents inherit reasoning effort and derive their own
effective mode.
- Retain the experimental app-server `multiAgentMode` fields for wire
compatibility while marking them deprecated. Request values are accepted
but ignored; compatibility response fields report `explicitRequestOnly`.
- Display Ultra in the TUI using the order supplied by `model/list`.
## Validation
- `just test -p codex-core ultra_reasoning_uses_max_for_requests`
- `just test -p codex-tui model_reasoning_selection_popup`
## Why
`WorldState` currently remembers its model-visible diff baseline only in
memory. That leaves no durable source for restoring the exact baseline
after resume, fork, rollback, or compaction.
This is the second PR in the WorldState persistence stack, built on
#29833 and following #29249. It records durable state transitions; the
next PR will replay them during rollout reconstruction.
## What
- Add a `world_state` rollout item containing either a full snapshot or
an RFC 7386 JSON Merge Patch.
- Persist a full snapshot after initial context and after compaction
establishes a new context window.
- Persist non-empty patches when later sampling steps or turns advance
the WorldState baseline.
- Write model-visible history before its matching WorldState record, so
an interrupted write can only cause a safe repeated update on replay.
- Preserve WorldState records for full-history forks while excluding
them from thread previews, metadata, and app-server history
materialization.
Older binaries read rollout lines independently, so they skip the
unknown `world_state` records while retaining the rest of the thread.
## Testing
- `just test -p codex-core
snapshot_merge_patch_changes_and_removes_nested_values`
- `just test -p codex-core
world_state_baseline_deduplicates_until_history_is_replaced`
- `just test -p codex-core
deferred_executor_compaction_preserves_then_updates_environment_once`
- `just test -p codex-protocol`
- `just test -p codex-rollout`
- `just test -p codex-state`
- `just test -p codex-thread-store`
- `just test -p codex-app-server-protocol`
## Why
`WorldState` currently keeps its diff baseline as live Rust objects
keyed by process-local `TypeId`. That baseline cannot be written to a
rollout or restored after resume, so Codex reconstructs an approximation
from `TurnContextItem`.
This is the first change in the WorldState persistence stack. It gives
every section a stable persisted identity and a compact serializable
comparison snapshot without changing rollout behavior yet.
## What changed
- Require each `WorldStateSection` to define a stable ID and
serializable snapshot type.
- Reject duplicate section IDs when constructing `WorldState`.
- Persist a dedicated environment comparison snapshot using
model-visible strings instead of runtime path types.
- Store only `WorldStateSnapshot` in `ContextManager`, removing the
parallel live-object baseline.
- Render diffs by restoring each section's typed snapshot; invalid
snapshots fall back to a full section render.
- Omit null object fields for future RFC 7386 patches while preserving
null values inside arrays.
Follow-up PRs will record full snapshots and merge patches, then restore
the baseline during resume, fork, and rollback.
## Test plan
- WorldState snapshot tests cover stable IDs, duplicate rejection, null
omission, and array preservation.
- Environment tests cover persistence-safe snapshot values and existing
diff rendering.
- ContextManager baseline deduplication and session context-update
persistence tests.
Related: #29249
## Why
Follow-up to #29249 and its [compaction review
thread](https://github.com/openai/codex/pull/29249#discussion_r3455055101).
During a turn, environment readiness can change between sampling
requests. Inline compaction must render the same model-visible
`WorldState` used by the request it follows. Rebuilding that state
during compaction can observe a newer environment, make replacement
history disagree with what the model saw, and suppress the next
environment update.
## What changed
- Make `run_turn` own the current `Arc<WorldState>` and replace it only
between sampling requests.
- Build each state from an explicitly chosen environment snapshot, diff
deferred-executor steps against the turn-owned state, and retain the
latest state in `ContextManager` only for cross-turn and resume
tracking.
- Pass the exact turn-owned state into inline compaction and explicit
new-context-window replacement.
- Carry that state with
`InitialContextInjection::BeforeLastUserMessage`, so replacement context
and its stored baseline cannot come from different snapshots.
- Remove obsolete state-recapture helpers and ambiguous TurnContext-only
WorldState builders.
- Add an integration test that moves an environment from starting to
ready during a paused turn, triggers compaction, and verifies the next
request receives the readiness update exactly once.
## Test plan
- `just test -p codex-core
deferred_executor_compaction_preserves_then_updates_environment_once`
- `just test -p codex-core process_compacted_history`
- `just test -p codex-core mid_turn_continuation_compaction`
- `just test -p codex-core build_initial_context`
- `just test -p codex-core
ignores_session_prefix_messages_when_truncating`
When using Responses Lite, we should all use `additional_tools` and a
developer item instead of the top level tools array & instructions
field. This keeps things 1-to-1.
Forced namespacing for _all_ tools will land in a following PR after
some coordination & fixes in Responses API (around collisions & return
items).
The goal is to eventually expand the scope of this to _all_ requests
from codex, but that will require larger coordination across providers &
slower rollout.
## Why
Environment context is model-visible state, but it is currently
assembled from transient turn values and diffed through
environment-specific paths. That makes initial injection, turn-to-turn
updates, and changes that happen within a turn use different baselines.
This PR introduces the smallest useful model world-state slice:
environments only, with one in-memory baseline and one renderer for full
state and diffs.
## What changed
- Add a typed `WorldState` container whose sections render fragments
relative to an optional previous value. Full rendering uses the same
diff path with no previous state.
- Replace the parallel `EnvironmentContext` representation with an
`EnvironmentsState` section keyed by environment ID and rendered in
deterministic order.
- Preserve the legacy single-environment output while supporting
multiple environments, starting environments, unavailable tombstones,
and changes to persisted turn-context values.
- Store the latest complete `WorldState` on `ContextManager` and use it
for both turn-boundary and mid-turn environment diffs.
- Build initial and post-compaction context from the same world-state
builder, then retain the rendered state as the next baseline.
- Seed the in-memory baseline from the latest `TurnContextItem` when
resuming an existing rollout; the world state itself is not serialized.
- Keep non-world settings updates on their existing path and merge
rendered world-state fragments at the session consumer.
## Known limitation
A legacy `TurnContextItem` only reconstructs the primary environment as
`local`; it cannot faithfully recover a remote-primary environment ID
after resume. Live state uses the exact environment IDs once a complete
baseline is established.
## Test plan
- `just test -p codex-core world_state`
- `just test -p codex-core record_context_updates`
- `just test -p codex-core deferred_executor_`
- `just test -p codex-core build_initial_context`
- `just test -p codex-core rollout_reconstruction`
- `just test -p codex-core
process_compacted_history_reinjects_full_initial_context`
## Description
This PR cuts Codex over from generic `ResponseItem.metadata` (introduced
here: https://github.com/openai/codex/pull/28355) to
`ResponseItem.internal_chat_message_metadata_passthrough`, which is the
blessed path and has strongly-typed keys.
For now we have to drop this MAv2 usage of `metadata`:
https://github.com/openai/codex/pull/28561 until we figure out where
that should live.
## Why
Nonblocking environment snapshots allow a turn to reach the model while
a remote environment is still starting. The initial context can describe
that environment as still loading, but nothing currently refreshes the
model-visible environment context when startup finishes during the same
turn.
This adds the first request-scoped reconciliation slice on top of
#28683. It is gated by `DeferredExecutor` and intentionally updates only
model-visible environment context; tools and other environment-derived
state will migrate separately.
## What
- Add a minimal `StepContext` containing the environment snapshot
captured before each sampling request.
- Render attached environments with their resolved shell and starting
environments with `still loading`.
- Track the latest environment state recorded in model history and
append a bounded update only when it changes.
- Seed that baseline from full initial context so ready-at-start
environments are not duplicated.
- Clear the in-memory baseline when history is rewritten so replacement
history can be refreshed safely.
## Testing
- `just test -p codex-core deferred_executor`
- `just test -p codex-core
environment_context_baseline_deduplicates_until_history_is_replaced`
The integration coverage verifies that a pending environment reaches the
first request, the ready state reaches the next request, later requests
do not duplicate it, and ready-at-start environments remain
single-injected.
<details>
<summary>Live verification</summary>
- Connected to a real remote executor with startup deliberately delayed
and forced three sampling requests in one turn.
- Inspected the raw model inputs: request 1 showed the remote
environment as `still loading`, request 2 appended its ready shell and
cwd, and request 3 contained no duplicate ready update.
- With the feature disabled, startup waited for the delayed executor and
the first request contained only the ready environment.
- With a synchronously ready environment and the feature enabled, the
first request contained one environment context with no duplicate.
- Executed `pwd` and read a marker file through the remote process
runner; the command exited successfully and returned the remote cwd and
marker contents.
</details>
## Why
Multi-agent delegation policy was split across `multiAgentMode`,
`features.multi_agent_mode`, and `usage_hint_enabled`. These controls
could disagree: a requested mode could be downgraded by the feature
flag, and disabling usage hints also disabled mode instructions.
Some clients also need multi-agent tools without adding
delegation-policy text to model context. The previous two-mode API could
not express that directly.
## What changed
`multiAgentMode` is now the only live delegation-policy control:
| Mode | Behavior |
| --- | --- |
| `none` | Keep multi-agent tools available without adding mode
instructions. |
| `explicitRequestOnly` | Only delegate after an explicit user request.
|
| `proactive` | Delegate when parallel work materially improves speed or
quality. |
- new threads default to `explicitRequestOnly`; omitting the mode on
later turns keeps the current value
- thread start, resume, fork, and settings responses always report the
concrete current mode instead of `null`
- mode selection remains sticky across turns and resume
- usage-hint text no longer controls whether mode instructions apply
- `features.multi_agent_mode` and `usage_hint_enabled` remain accepted
as ignored compatibility settings so existing configs continue to load
- app-server documentation and generated schemas describe the three-mode
API
## Tests
- `just test -p codex-core multi_agent_mode`
- `just test -p codex-core multi_agent_v2_config_from_feature_table`
- `just test -p codex-core spawn_agent_description`
- `just test -p codex-features`
- `just test -p codex-app-server-protocol`
- `just test -p codex-app-server multi_agent_mode`
## Why
Once multi-agent mode can be selected per turn, clients also need to
choose the initial selection when creating a thread and observe that
selection through lifecycle and settings APIs.
The selected value is intentionally distinct from the effective
model-visible value: no client selection is represented as `null`, even
though an eligible multi-agent v2 turn derives `explicitRequestOnly` as
its effective default.
## What changed
- Add the optional experimental `thread/start.multiAgentMode` parameter
and pass it through thread creation.
- Preserve an omitted initial value as an unset selection rather than
eagerly storing `explicitRequestOnly`.
- Apply an explicit `thread/start` selection to the first turn through
the session configuration established at thread creation.
- Restore the latest persisted effective mode as the selected baseline
on cold resume when rollout history contains one.
- Inherit the optional selected mode from a loaded parent when creating
related runtime threads.
- Return the current selected `multiAgentMode` from `thread/start`,
`thread/resume`, `thread/fork`, and thread settings, using `null` when
no mode is selected.
- Keep lifecycle reporting independent from model capability and feature
eligibility; core turn construction remains responsible for calculating
and persisting the effective mode.
## Not covered
- Clearing an existing loaded-session selection back to unset through
`turn/start`; omitted or `null` currently retains the session's
selection.
- A TUI control, slash command, or `config.toml` preference.
## Verification
- `CARGO_INCREMENTAL=0 just test -p codex-app-server-protocol`
- `CARGO_INCREMENTAL=0 just test -p codex-app-server multi_agent_mode`
The focused app-server coverage verifies explicit `thread/start`
initialization, first-turn prompting, nullable reporting for an omitted
selection, and retention of selections that are not currently
runtime-eligible.
## Stack
Stacked on #28685. This PR contains only the thread initialization and
lifecycle/settings API layer.
## Why
Multi-agent v2 currently carries an explicit-request-only delegation
rule in its static usage hint. That provides a safe default, but it
prevents clients from selecting proactive delegation per turn without
changing static guidance or rewriting prior model context.
This change makes delegation mode a session selection that can be
updated through `turn/start`, while deriving the effective model-visible
mode separately for each turn. Eligible multi-agent v2 turns remain
explicit-request-only unless proactive mode is both selected and
enabled.
## What changed
- Add the experimental `turn/start.multiAgentMode` parameter with
`explicitRequestOnly` and `proactive` values. Omission retains the
loaded session's current optional selection.
- Add the default-off `features.multi_agent_mode` feature gate. Eligible
multi-agent v2 turns use the selected mode when enabled; an unset
selection or disabled gate resolves to `explicitRequestOnly`.
- Treat mode prompting as inapplicable for multi-agent v1 and other
unsupported session configurations, producing no multi-agent mode
developer message rather than rejecting the turn.
- Move the explicit-request-only rule out of the static v2 usage hint
and into a bounded, tagged developer context fragment.
- Emit the effective mode in initial context and only when that
effective mode changes on later turns.
- Persist the effective mode in `TurnContextItem` as the durable
baseline for resume and context-update comparisons.
Historical rollout items are not rewritten. Later mode developer
messages establish the current rule incrementally.
## Not covered
- Initial selection through `thread/start` and selected-mode reporting
from thread lifecycle/settings APIs; those are isolated in the stacked
#28792.
- A TUI control or slash command for selecting the mode.
- Persisting a preferred mode to `config.toml`; selection remains
session/turn scoped.
- Changes to multi-agent concurrency limits, tool availability, or model
catalog capability declarations.
- Rewriting historical rollout prompt items. Cold resume restores the
latest persisted effective mode when available while leaving historical
developer messages intact.
## Verification
- `CARGO_INCREMENTAL=0 just test -p codex-core multi_agent_mode`
- Focused app-server coverage verifies that `turn/start.multiAgentMode`
produces proactive developer instructions for an eligible v2 turn.
## Stack
Followed by #28792, which adds `thread/start` initialization and
lifecycle/settings observability.
## Why
`ResponseItem` variants do not have a consistent internal ID shape: some
variants carry required IDs, some carry optional IDs, and some cannot
represent an ID at all. The existing fields also use inconsistent serde,
TypeScript, and JSON-schema annotations. A single enum-level access path
is needed before history recording can assign and retain IDs.
This PR establishes that internal model only. It intentionally does not
generate or serialize IDs; allocation and wire persistence are isolated
in the stacked follow-up.
## What changed
- Give every concrete `ResponseItem` variant an `Option<String>` ID
field.
- Apply the same internal-only annotations to every ID field:
`#[serde(default, skip_serializing)]`, `#[ts(skip)]`, and
`#[schemars(skip)]`.
- Add `ResponseItem::id()` and `ResponseItem::set_id()` as the shared
accessors.
- Preserve IDs when history items are rewritten for truncation.
- Adapt consumers that previously assumed reasoning and image-generation
IDs were required.
- Regenerate app-server schemas so the hidden fields are represented
consistently.
The serde catch-all `ResponseItem::Other` remains ID-less because it
must remain a unit variant.
## Test plan
- `cargo check --tests -p codex-core -p codex-api -p codex-rollout-trace
-p codex-image-generation-extension`
- `just test -p codex-protocol`
- `just test -p codex-app-server-protocol`
- `just test -p codex-api -p codex-rollout-trace -p
codex-image-generation-extension`
- `just test -p codex-core event_mapping`
## Summary
This keeps inter-agent communication on the existing raw response item
path and adds a join key for MAv2 tool calls.
MAv2 `spawn_agent`, `send_message`, and `followup_task` now stamp the
originating tool call id into `ResponseItemMetadata.source_call_id` on
the raw `ResponseItem::AgentMessage`. App-server clients can join that
raw item back to the existing tool/activity event by call id, while
using the raw agent message's existing sender, receiver, and content
fields.
No new app-server `ThreadItem` or notification type is added.
## Tests
- `just fmt`
- `just write-app-server-schema`
- `just test -p codex-protocol`
- `just test -p codex-app-server-protocol`
- `just test -p codex-core
multi_agent_v2_spawn_returns_path_and_send_message_accepts_relative_path`
- `just test -p codex-core
multi_agent_v2_followup_task_completion_notifies_parent_on_every_turn`
- `just fix -p codex-protocol`
- `just fix -p codex-app-server-protocol`
- `just fix -p codex-core`
## Why
`TurnContext` had accumulated dead fields and cached projections of
values already owned by its per-turn `Config` or `ModelInfo`. Keeping
both copies made ownership unclear and allowed artificial split-brain
states, such as a compatibility hash differing from the model metadata
it came from.
`Prompt` similarly carried a write-only personality after personality
selection had already been materialized into its base instructions.
This makes the canonical owner explicit: configuration-backed values
come from `config`, model-derived values come from `model_info`, and
prompts contain only data consumed by request construction.
## What changed
- Remove the unused `ghost_snapshot`, `codex_self_exe`, and
`thread_source` fields.
- Remove duplicate `comp_hash`, `truncation_policy`, `features`,
`shell_environment_policy`, `codex_linux_sandbox_exe`, `compact_prompt`,
and `tool_mode` fields.
- Read those values directly from `TurnContext::config` or
`TurnContext::model_info` at their consumers.
- Remove the write-only `Prompt::personality` field and its constructor
assignments.
- Preserve review-turn inheritance of the parent turn's shell policy,
Linux sandbox executable, and compact prompt through the review config.
## Testing
- `cargo check -p codex-core --tests`
## Why
#28152 jumped the gun on moving the rollout format to store URIs, and
would likely break compat with some features that don't go through the
same types as the core logic.
## What
Make `TurnContextItem.cwd` an `AbsolutePathBuf` again, remove test added
for `PathUri` serialization in rollouts. Also drops a bunch of error
paths that are no longer needed.
## Why
Model-visible `<environment_context>` should match the environment of
the executor, not of the app server.
Stacked on #28146.
## What
- Keep selected environment cwd values as `PathUri` while building
environment context.
- Render cwd text using the path convention represented by the URI, with
the canonical URI as a fallback.
- Preserve compatibility with legacy `TurnContextItem.cwd` values when
reconstructing and diffing context.
- Extend the Wine-backed remote Windows test to assert that the model
sees `powershell` and `C:\windows`.
## Description
This PR adds an optional `metadata` field to `ResponseItem` for
Responses API calls. Only mechanical plumbing, no actual values
populated and sent yet. Turns out just adding a new field to
`ResponseItem` has quite a large blast radius already.
This change is backwards compatible because `metadata` is optional and
omitted when absent, so existing response items and rollout history
without it still deserialize and requests that do not set it keep the
same wire shape. For provider compatibility, we strip out `metadata`
before non-OpenAI Responses requests so Azure and AWS Bedrock never see
this field.
My followup PR here will actually make use of it to start storing and
passing along `turn_id`: https://github.com/openai/codex/pull/28360
## What changed
- Added `ResponseItemMetadata` with optional `turn_id`, plus optional
`metadata` on Responses API item variants and inter-agent communication.
- Preserved item metadata through response-item rewrites such as
truncation, missing tool-output synthesis, compaction history
rebuilding, visible-history conversion, rollout/resume, and generated
app-server schemas/types.
- Strip item metadata from non-OpenAI Responses requests while
preserving it for OpenAI-shaped requests.
- Updated the mechanical fixture/test construction churn required by the
new optional field.
## Why
When we prepare the conversation history, every tool call needs a
matching output.
Before this change, we scanned the full history again for every call. In
a tool-heavy conversation, that makes the work `O(items x calls)`, or
`O(n^2)` in the worst case.
## What
Scan the history once and collect the IDs of existing outputs. Then each
call can check its ID with an expected `O(1)` lookup.
The full normalization step is now expected `O(n)`. The output order and
missing-output behavior stay the same.
## Performance
Based on local rollout traces, one tool-heavy session reached roughly
17,050 transcript items with about 4,292 tool-call items. On a history
of that shape, the old `calls x items` scan does about 73.2 million
membership checks, while the new pass does about 21.3 thousand set
inserts/lookups. That is roughly 3.4k times less membership work in this
normalization step.
## Validation
- `just test -p codex-core normalize_` (19 passed)
## Summary
- snapshot `comp_hash` into `TurnContext` when the turn is created and
use that snapshot as the downstream source of truth
- persist the turn hash in rollout context and recover it into
previous-turn settings during resume and fork replay
- compact existing history with the previous model only when both
adjacent turns provide hashes and the values differ
- record `comp_hash_changed` as the compaction reason
- cover ordinary transitions, resume, and missing-hash compatibility
with end-to-end tests
## Why
History produced under one compaction-compatible model configuration may
not be safe to carry directly into another. Compacting at the turn
boundary converts that history before context updates and the new user
message are added. Persisting the turn snapshot in `TurnContextItem`
makes the same protection work after resuming a rollout.
A missing hash is not treated as evidence of incompatibility. `None →
Some`, `Some → None`, and `None → None` do not trigger compaction; only
`Some(previous) → Some(current)` with unequal values does.
## Stack
- depends on #27532
- #27532 is based directly on `main`
## Testing
- `just test -p codex-core pre_sampling_compact_` — 6 passed
- `just test -p codex-core
turn_context_item_uses_turn_context_comp_hash_snapshot` — passed
- `just fix -p codex-core -p codex-protocol -p codex-analytics -p
codex-models-manager`
## Why
`log_remote_compact_failure` was the only consumer of the
compact-request logging payload and most of the token-usage breakdown
fields. Once that failure log is removed, keeping the surrounding
carrier types leaves dead plumbing in the compaction path and context
manager.
## What changed
- Remove `log_remote_compact_failure`, `CompactRequestLogData`, and the
v2 wrapper that only fed that log.
- Let both remote compaction implementations return the original
compaction error directly.
- Replace `TotalTokenUsageBreakdown` with a narrow helper that returns
only the remaining value needed by compaction analytics.
- Keep `estimate_response_item_model_visible_bytes` private to the
context manager implementation.
## Validation
- `cargo check -p codex-core`
## Why
Multi-agent v2 currently routes agent instructions through normal tool
arguments and inter-agent context. That means the parent model can emit
plaintext task text, Codex can persist it in history/rollouts, and the
recipient can receive it as ordinary assistant-message JSON.
This changes the v2 path so agent instructions stay encrypted between
model calls: Responses encrypts the `message` argument returned by the
model, Codex forwards only that ciphertext, and Responses decrypts it
internally for the recipient model.
## What changed
- Mark the v2 `message` parameter as encrypted for `spawn_agent`,
`send_message`, and `followup_task`.
- Treat multi-agent v2 tool `message` values as ciphertext
unconditionally.
- Store v2 inter-agent task text in
`InterAgentCommunication.encrypted_content` with empty plaintext
`content`.
- Convert encrypted inter-agent communications into the Responses
`agent_message` input item before sending the child request.
- Preserve `agent_message` items across history, rollout, compaction,
telemetry, and app-server schema paths.
- Leave multi-agent v1 unchanged.
## Message shape
The model still calls the v2 tools with a `message` argument, but that
value is now ciphertext:
```json
{
"name": "spawn_agent",
"arguments": {
"task_name": "worker",
"message": "<ciphertext>"
}
}
```
Codex stores the task as encrypted inter-agent communication:
```json
{
"author": "/root",
"recipient": "/root/worker",
"content": "",
"encrypted_content": "<ciphertext>",
"trigger_turn": true
}
```
When Codex builds the recipient request, it forwards the ciphertext
using the new Responses input item:
```json
{
"type": "agent_message",
"author": "/root",
"recipient": "/root/worker",
"content": [
{
"type": "encrypted_content",
"encrypted_content": "<ciphertext>"
}
]
}
```
Responses decrypts that item internally for the recipient model.
## Context impact
- Parent context no longer carries plaintext v2 agent task instructions
from these tool arguments.
- Codex rollout/history stores ciphertext for v2 agent instructions.
- Recipient requests receive an `agent_message` item instead of
assistant commentary JSON for encrypted task delivery.
- Plaintext completion/status notifications are still plaintext because
they are Codex-generated status messages, not encrypted model tool
arguments.
## Validation
- `just test -p codex-tools`
- `just test -p codex-protocol`
- `just test -p codex-rollout`
- `just test -p codex-rollout-trace`
- `just test -p codex-otel`
- `just write-app-server-schema`
