Fixes#11845.
Adjust context/token estimation for inline image `data:*;base64,...`
URLs so we
do not count the raw base64 payload as model-visible text.
What changed:
- keep the existing JSON-length estimator as the baseline
- detect only inline base64 `data:` image URLs in message and
function-call
output content items
- subtract only the base64 payload bytes (preserving data URL prefix +
JSON
overhead)
- add a fixed per-image estimate of 340 bytes (~85 tokens at the repo’s
4-bytes/token heuristic)
This avoids large overestimates from MCP image tool outputs while
leaving normal
image URLs (`https://`, `file://`, non-base64 `data:` URLs) unchanged.
Tests:
- message image data URL estimate regression
- function-call output image data URL estimate regression
- non-base64 image URLs unchanged
- non-base64 `data:` URLs unchanged
- `data:application/octet-stream;base64,...` adjusted
- multiple inline images apply multiple fixed costs
- text-only items unchanged
## Why
`codex-core` was carrying the embedded system-skill sample assets (and a
`build.rs` that walks those files to register rerun triggers). Those
assets change infrequently, but any change under `codex-core` still ties
them to `codex-core`'s build/cache lifecycle.
This change moves the embedded system-skills packaging into a dedicated
`codex-skills` crate so it can be cached independently. That reduces
unnecessary invalidation/rebuild pressure on `codex-core` when the
skills bundle is the only thing that changes.
## What Changed
- Added a new `codex-rs/skills` crate (`codex-skills`) with:
- `Cargo.toml`
- `BUILD.bazel`
- `build.rs` to track skill asset file changes for Cargo rebuilds
- `src/lib.rs` containing the embedded system-skills install/cache logic
previously in `codex-core`
- Moved the embedded sample skill assets from
`codex-rs/core/src/skills/assets/samples` to
`codex-rs/skills/src/assets/samples`.
- Updated `codex-rs/core/Cargo.toml` to depend on `codex-skills` and
removed `codex-core`'s direct `include_dir` dependency.
- Removed `codex-core`'s `build.rs`.
- Replaced `codex-rs/core/src/skills/system.rs` implementation with a
thin re-export wrapper to keep existing `codex-core` call sites
unchanged.
- Updated workspace manifests/lockfile (`codex-rs/Cargo.toml`,
`codex-rs/Cargo.lock`) for the new crate.
## Why
`codex-rs/arg0` only needed two things from `codex-core`:
- the `find_codex_home()` wrapper
- the special argv flag used for the internal `apply_patch`
self-invocation path
That made `codex-arg0` depend on `codex-core` for a very small surface
area. This change removes that dependency edge and moves the shared
`apply_patch` invocation flag to a more natural boundary
(`codex-apply-patch`) while keeping the contract explicitly documented.
## What Changed
- Moved the internal `apply_patch` argv[1] flag constant out of
`codex-core` and into `codex-apply-patch`.
- Renamed the constant to `CODEX_CORE_APPLY_PATCH_ARG1` and documented
that it is part of the Codex core process-invocation contract (even
though it now lives in `codex-apply-patch`).
- Updated `arg0`, the core apply-patch runtime, and the `codex-exec`
apply-patch test to import the constant from `codex-apply-patch`.
- Updated `codex-rs/arg0` to call
`codex_utils_home_dir::find_codex_home()` directly instead of
`codex_core::config::find_codex_home()`.
- Removed the `codex-core` dependency from `codex-rs/arg0` and added the
needed direct dependency on `codex-utils-home-dir`.
- Added `codex-apply-patch` as a dev-dependency for `codex-rs/exec`
tests (the apply-patch test now imports the moved constant directly).
## Verification
- `cargo test -p codex-apply-patch`
- `cargo test -p codex-arg0`
- `cargo test -p codex-core --lib apply_patch`
- `cargo test -p codex-exec
test_standalone_exec_cli_can_use_apply_patch`
- `cargo shear`
## Why
`codex-rs/core/src/lib.rs` re-exported a broad set of types and modules
from `codex-protocol` and `codex-shell-command`. That made it easy for
workspace crates to import those APIs through `codex-core`, which in
turn hides dependency edges and makes it harder to reduce compile-time
coupling over time.
This change removes those public re-exports so call sites must import
from the source crates directly. Even when a crate still depends on
`codex-core` today, this makes dependency boundaries explicit and
unblocks future work to drop `codex-core` dependencies where possible.
## What Changed
- Removed public re-exports from `codex-rs/core/src/lib.rs` for:
- `codex_protocol::protocol` and related protocol/model types (including
`InitialHistory`)
- `codex_protocol::config_types` (`protocol_config_types`)
- `codex_shell_command::{bash, is_dangerous_command, is_safe_command,
parse_command, powershell}`
- Migrated workspace Rust call sites to import directly from:
- `codex_protocol::protocol`
- `codex_protocol::config_types`
- `codex_protocol::models`
- `codex_shell_command`
- Added explicit `Cargo.toml` dependencies (`codex-protocol` /
`codex-shell-command`) in crates that now import those crates directly.
- Kept `codex-core` internal modules compiling by using `pub(crate)`
aliases in `core/src/lib.rs` (internal-only, not part of the public
API).
- Updated the two utility crates that can already drop a `codex-core`
dependency edge entirely:
- `codex-utils-approval-presets`
- `codex-utils-cli`
## Verification
- `cargo test -p codex-utils-approval-presets`
- `cargo test -p codex-utils-cli`
- `cargo check --workspace --all-targets`
- `just clippy`
## Why
Compiling `codex-rs/core` is a bottleneck for local iteration, so this
change continues the ongoing extraction of config-related functionality
out of `codex-core` and into `codex-config`.
The goal is not just to move code, but to reduce `codex-core` ownership
and indirection so more code depends on `codex-config` directly.
## What Changed
- Moved config diagnostics logic from
`core/src/config_loader/diagnostics.rs` into
`config/src/diagnostics.rs`.
- Updated `codex-core` to use `codex-config` diagnostics types/functions
directly where possible.
- Removed the `core/src/config_loader/diagnostics.rs` shim module
entirely; the remaining `ConfigToml`-specific calls are in
`core/src/config_loader/mod.rs`.
- Moved `CONFIG_TOML_FILE` into `codex-config` and updated existing
references to use `codex_config::CONFIG_TOML_FILE` directly.
- Added a direct `codex-config` dependency to `codex-cli` for its
`CONFIG_TOML_FILE` use.
## Summary
- move regular-turn context diff/full-context persistence into
`run_turn` so pre-turn compaction runs before incoming context updates
are recorded
- after successful pre-turn compaction, rely on a cleared
`reference_context_item` to trigger full context reinjection on the
follow-up regular turn (manual `/compact` keeps replacement history
summary-only and also clears the baseline)
- preserve `<model_switch>` when full context is reinjected, and inject
it *before* the rest of the full-context items
- scope `reference_context_item` and `previous_model` to regular user
turns only so standalone tasks (`/compact`, shell, review, undo) cannot
suppress future reinjection or `<model_switch>` behavior
- make context-diff persistence + `reference_context_item` updates
explicit in the regular-turn path, with clearer docs/comments around the
invariant
- stop persisting local `/compact` `RolloutItem::TurnContext` snapshots
(only regular turns persist `TurnContextItem` now)
- simplify resume/fork previous-model/reference-baseline hydration by
looking up the last surviving turn context from rollout lifecycle
events, including rollback and compaction-crossing handling
- remove the legacy fallback that guessed from bare `TurnContext`
rollouts without lifecycle events
- update compaction/remote-compaction/model-visible snapshots and
compact test assertions (including remote compaction mock response
shape)
## Why
We were persisting incoming context items before spawning the regular
turn task, which let pre-turn compaction requests accidentally include
incoming context diffs without the new user message. Fixing that exposed
follow-on baseline issues around `/compact`, resume/fork, and standalone
tasks that could cause duplicate context injection or suppress
`<model_switch>` instructions.
This PR re-centers the invariants around regular turns:
- regular turns persist model-visible context diffs/full reinjection and
update the `reference_context_item`
- standalone tasks do not advance those regular-turn baselines
- compaction clears the baseline when replacement history may have
stripped the referenced context diffs
## Follow-ups (TODOs left in code)
- `TODO(ccunningham)`: fix rollback/backtracking baseline handling more
comprehensively
- `TODO(ccunningham)`: include pending incoming context items in
pre-turn compaction threshold estimation
- `TODO(ccunningham)`: inject updated personality spec alongside
`<model_switch>` so some model-switch paths can avoid forced full
reinjection
- `TODO(ccunningham)`: review task turn lifecycle
(`TurnStarted`/`TurnComplete`) behavior and emit task-start context
diffs for task types that should have them (excluding `/compact`)
## Validation
- `just fmt`
- CI should cover the updated compaction/resume/model-visible snapshot
expectations and rollout-hydration behavior
- I did **not** rerun the full local test suite after the latest
resume-lookup / rollout-persistence simplifications
## Summary
`gpt-5.3-codex` really likes to write complicated shell scripts, and
suggest a partial prefix_rule that wouldn't actually approve the
command. We should only show the `prefix_rule` suggestion from the model
if it would actually fully approve the command the user is seeing.
This will technically cause more instances of overly-specific
suggestions when we fallback, but I think the UX is clearer,
particularly when the model doesn't necessarily understand the current
limitations of execpolicy parsing.
## Testing
- [x] Add unit tests
- [x] Add integration tests
- add top-level `experimental_realtime_ws_backend_prompt` config key
(experimental / do not use) and include it in config schema
- apply the override only to `Op::RealtimeConversation` websocket
`backend_prompt`, with config + realtime tests
Addresses https://github.com/openai/codex/issues/11013
## Summary
- add a Plan implementation path in the TUI that lets users choose
reasoning before switching to Default mode and implementing
- add Plan-mode reasoning scope handling (Plan-only override vs
all-modes default), including config/schema/docs plumbing for
`plan_mode_reasoning_effort`
- remove the hardcoded Plan preset medium default and make the reasoning
popup reflect the active Plan override as `(current)`
- split the collaboration-mode switch notification UI hint into #12307
to keep this diff focused
If I have `plan_mode_reasoning_effort = "medium"` set in my
`config.toml`:
<img width="699" height="127" alt="Screenshot 2026-02-20 at 6 59 37 PM"
src="https://github.com/user-attachments/assets/b33abf04-6b7a-49ed-b2e9-d24b99795369"
/>
If I don't have `plan_mode_reasoning_effort` set in my `config.toml`:
<img width="704" height="129" alt="Screenshot 2026-02-20 at 7 01 51 PM"
src="https://github.com/user-attachments/assets/88a086d4-d2f1-49c7-8be4-f6f0c0fa1b8d"
/>
## Codex author
`codex resume 019c78a2-726b-7fe3-adac-3fa4523dcc2a`
- add top-level `experimental_realtime_ws_base_url` config key
(experimental / do not use) and include it in config schema
- apply the override only to `Op::RealtimeConversation` websocket
transport, with config + realtime tests
- Introduce `RealtimeConversationManager` for realtime API management
- Add `op::conversation` to start conversation, insert audio, insert
text, and close conversation.
- emit conversation lifecycle and realtime events.
- Move shared realtime payload types into codex-protocol and add core
e2e websocket tests for start/replace/transport-close paths.
Things to consider:
- Should we use the same `op::` and `Events` channel to carry audio? I
think we should try this simple approach and later we can create
separate one if the channels got congested.
- Sending text updates to the client: we can start simple and later
restrict that.
- Provider auth isn't wired for now intentionally
# External (non-OpenAI) Pull Request Requirements
Before opening this Pull Request, please read the dedicated
"Contributing" markdown file or your PR may be closed:
https://github.com/openai/codex/blob/main/docs/contributing.md
If your PR conforms to our contribution guidelines, replace this text
with a detailed and high quality description of your changes.
Include a link to a bug report or enhancement request.
## Summary
- switch the review test SSE mock helper to use the shared hermetic mock
server setup
- ensure review tests always have a default `/v1/models` stub during
Codex session bootstrap
- remove the race that caused intermittent `/v1/models` connection
failures and flaky ETag refresh assertions
## Testing
- `just fmt`
- `cargo test -p codex-core --test all
refresh_models_on_models_etag_mismatch_and_avoid_duplicate_models_fetch`
- `cargo test -p codex-core --test all
review_uses_custom_review_model_from_config`
- repeated both targeted tests 5x in a loop
- `cargo clippy -p codex-core --tests -- -D warnings`
Summary
- ensure destructive tool annotations short-circuit to require approval
- simplify approval logic to only require read/write + open-world when
destructive is false
- update the unit test to cover the new destructive behavior
Testing
- Not run (not requested)
## Summary
Adds support for a Unix socket escape hatch so we can bypass socket
allowlisting when explicitly enabled.
## Description
* added a new flag, `network.dangerously_allow_all_unix_sockets` as an
explicit escape hatch
* In codex-network-proxy, enabling that flag now allows any absolute
Unix socket path from x-unix-socket instead of requiring each path to be
explicitly allowlisted. Relative paths are still rejected.
* updated the macOS seatbelt path in core so it enforces the same Unix
socket behavior:
* allowlisted sockets generate explicit network* subpath rules
* allow-all generates a broad network* (subpath "/") rule
---------
Co-authored-by: Codex <199175422+chatgpt-codex-connector[bot]@users.noreply.github.com>
## Summary
Tighten the `js_repl` freeform Lark grammar to block the most common
malformed payload wrappers before they reach runtime validation.
## What Changed
- Replaced the overly permissive `js_repl` freeform grammar (`start:
/[\s\S]*/`) with a structured grammar that still supports:
- plain JS source
- optional first-line `// codex-js-repl:` pragma followed by JS source
- Added grammar-level filtering for common bad payload shapes by
rejecting inputs whose first significant token starts with:
- `{` (JSON object wrapper like `{"code":"..."}`)
- `"` (quoted code string)
- `` ``` `` (markdown code fences)
- Implemented the grammar without regex lookahead/lookbehind because the
API-side Lark regex engine does not support look-around.
- Added a unit test to validate the grammar shape and guard against
reintroducing unsupported lookaround.
## Why
`js_repl` is a freeform tool, but the model sometimes emits wrapped
payloads (JSON, quoted strings, markdown fences) instead of raw
JavaScript. We already reject those at runtime, but this change moves
the constraint into the tool grammar so the model is less likely to
generate invalid tool-call payloads in the first place.
## Testing
- `cargo test -p codex-core
js_repl_freeform_grammar_blocks_common_non_js_prefixes`
- `cargo test -p codex-core parse_freeform_args_rejects_`
## Notes
- This intentionally over-blocks a few uncommon valid JS starts (for
example top-level `{ ... }` blocks or top-level quoted directives like
`"use strict";`) in exchange for preventing the common wrapped-payload
mistakes.
#### [git stack](https://github.com/magus/git-stack-cli)
- 👉 `1` https://github.com/openai/codex/pull/12300
- ⏳ `2` https://github.com/openai/codex/pull/12275
- ⏳ `3` https://github.com/openai/codex/pull/12205
- ⏳ `4` https://github.com/openai/codex/pull/12185
- ⏳ `5` https://github.com/openai/codex/pull/10673
## Summary
Simplify network approvals by removing per-attempt proxy correlation and
moving to session-level approval dedupe keyed by (host, protocol, port).
Instead of encoding attempt IDs into proxy credentials/URLs, we now
treat approvals as a destination policy decision.
- Concurrent calls to the same destination share one approval prompt.
- Different destinations (or same host on different ports) get separate
prompts.
- Allow once approves the current queued request group only.
- Allow for session caches that (host, protocol, port) and auto-allows
future matching requests.
- Never policy continues to deny without prompting.
Example:
- 3 calls:
- a.com (line 443)
- b.com (line 443)
- a.com (line 443)
=> 2 prompts total (a, b), second a waits on the first decision.
- a.com:80 is treated separately from a.com line 443
## Testing
- `just fmt` (in `codex-rs`)
- `cargo test -p codex-core tools::network_approval::tests`
- `cargo test -p codex-core` (unit tests pass; existing
integration-suite failures remain in this environment)
Summary
- capture the origin for each configured MCP server and expose it via
the connection manager
- plumb MCP server name/origin into tool logging and emit
codex.tool_result events with those fields
- add unit coverage for origin parsing and extend OTEL tests to assert
empty MCP fields for non-MCP tools
- currently not logging full urls or url paths to prevent logging
potentially sensitive data
Testing
- Not run (not requested)
## Summary
- add reasoning effort constants for the memories phase one and phase
two agents
- wire the constants into phase1 request creation and phase2 agent
configuration so the default efforts are always applied
## Testing
- Not run (not requested)
## Summary
- Add `rollout_summary_file: <generated>.md` to each thread header in
`raw_memories.md` so Phase 2 can reliably reference the canonical
rollout summary filename.
- Update the memory prompts/templates (`stage_one_system`,
`consolidation`, `read_path`) for the new task-oriented raw-memory /
MEMORY.md schema and stronger consolidation guidance.
## Details
- `codex-rs/core/src/memories/storage.rs`
- Writes the generated `rollout_summary_file` path into the per-thread
metadata header when rebuilding `raw_memories.md`.
- `codex-rs/core/src/memories/tests.rs`
- Verifies the canonical `rollout_summary_file` header is present and
ordered after `updated_at`/`cwd` in `raw_memories.md`.
- Verifies task-structured raw-memory content is preserved while the
canonical header is added.
- `codex-rs/core/templates/memories/*.md`
- Updates the stage-1 raw-memory format to task-grouped sections
(`task`, `task_group`, `task_outcome`).
- Updates Phase 2 consolidation guidance around recency (`updated_at`),
task-oriented `MEMORY.md` blocks, and richer evidence-backed
consolidation.
- Tweaks the quick memory pass wording to emphasize topics/workflows in
addition to keywords.
## Testing
- `cargo test -p codex-core memories`
We now write MCP tools from installed apps to disk cache so that they
can be picked up instantly at startup. We still do a fresh fetch from
remote MCP server but it's non blocking unless there's a cache miss.
- [x] Store apps tool cache in disk to reduce startup time.
## Summary
Implements the `ConfigToml.permissions.network` and uses it to populate
`NetworkProxyConfig`. We now parse a new nested permissions/network
config shape which is converted into the proxy’s runtime config.
When managed requirements exist, we still apply those constraints on top
of user settings (so managed policy still wins).
* Cleaned up the old constructor path so it now accepts both user config
+ managed constraints directly.
* Updated the reload path so live proxy config reloads respect
[permissions.network] too, while still supporting the existing top-level
[network] format.
### Behavior
- User-defined `[permissions.network]` values are now honored.
- Managed constraints still take effect and are validated against the
resulting policy.
## Summary
Implements a configurable MCP OAuth callback URL override for `codex mcp
login` and app-server OAuth login flows, including support for non-local
callback endpoints (for example, devbox ingress URLs).
## What changed
- Added new config key: `mcp_oauth_callback_url` in
`~/.codex/config.toml`.
- OAuth authorization now uses `mcp_oauth_callback_url` as
`redirect_uri` when set.
- Callback handling validates the callback path against the configured
redirect URI path.
- Listener bind behavior is now host-aware:
- local callback URL hosts (`localhost`, `127.0.0.1`, `::1`) bind to
`127.0.0.1`
- non-local callback URL hosts bind to `0.0.0.0`
- `mcp_oauth_callback_port` remains supported and is used for the
listener port.
- Wired through:
- CLI MCP login flow
- App-server MCP OAuth login flow
- Skill dependency OAuth login flow
- Updated config schema and config tests.
## Why
Some environments need OAuth callbacks to land on a specific reachable
URL (for example ingress in remote devboxes), not loopback. This change
allows that while preserving local defaults for existing users.
## Backward compatibility
- No behavior change when `mcp_oauth_callback_url` is unset.
- Existing `mcp_oauth_callback_port` behavior remains intact.
- Local callback flows continue binding to loopback by default.
## Testing
- `cargo test -p codex-rmcp-client callback -- --nocapture`
- `cargo test -p codex-core --lib mcp_oauth_callback -- --nocapture`
- `cargo check -p codex-cli -p codex-app-server -p codex-rmcp-client`
## Example config
```toml
mcp_oauth_callback_port = 5555
mcp_oauth_callback_url = "https://<devbox>-<namespace>.gateway.<cluster>.internal.api.openai.org/callback"
## Summary
- The experimental Bazel CI builds fail on all platforms because askama
resolves template paths relative to `CARGO_MANIFEST_DIR`, which points
outside the Bazel sandbox. This produces errors like:
```
error: couldn't read
`codex-rs/core/src/memories/../../../../../../../../../../../work/codex/codex/codex-rs/core/templates/memories/consolidation.md`:
No such file or directory
```
- Replaced `#[derive(Template)]` + `#[template(path = "...")]` with
`include_str!` + `str::replace()` for the three affected templates
(`consolidation.md`, `stage_one_input.md`, `read_path.md`).
`include_str!` resolves paths relative to the source file, which works
correctly in both Cargo and Bazel builds.
- The templates only use simple `{{ variable }}` substitution with no
control flow or filters, so no askama functionality is lost.
- Removes the `askama` dependency from `codex-core` since it was the
only crate using it. The workspace-level dependency definition is left
in place.
- This matches the existing pattern used throughout the codebase — e.g.
`codex-rs/core/src/memories/mod.rs` already uses
`include_str!("../../templates/memories/stage_one_system.md")` for the
fourth template file.
## Test plan
- [ ] Verify Bazel (experimental) CI passes on all platforms
- [ ] Verify rust-ci (Cargo) builds and tests continue to pass
- [ ] Verify `cargo test -p codex-core` passes locally
## Summary
- Require revised `<proposed_plan>` blocks in the same planning session
to be complete replacements, not partial/delta plans.
- Scope that cumulative replacement rule to the current planning session
only.
- Clarify that after leaving Plan mode (for example switching to Default
mode to implement) or when explicitly asked for a new plan, the model
should produce a new self-contained plan without inheriting prior plan
blocks unless requested.
## Testing
- Not run (prompt/template text-only change).
Summary
- avoid emitting metrics for features marked as `Stage::Removed`
- keep feature metrics aligned with active and planned states only
Testing
- Not run (not requested)
## Why
`McpConnectionManager` used a two-phase setup (`new()` followed by
`initialize()`), which forced call sites to construct placeholder state
and then mutate it asynchronously. That made MCP startup/refresh flows
harder to follow and easier to misuse, especially around cancellation
token ownership.
## What changed
- Replaced the two-phase initialization flow with a single async
constructor: `McpConnectionManager::new(...) -> (Self,
CancellationToken)`.
- Added `McpConnectionManager::new_uninitialized()` for places that need
an empty manager before async startup begins.
- Added `McpConnectionManager::new_mcp_connection_manager_for_tests()`
for test-only construction.
- Updated MCP startup and refresh call sites in
`codex-rs/core/src/codex.rs` to build a fresh manager via `new(...)`,
swap it in, and update the startup cancellation token consistently.
- Updated MCP snapshot/connector call sites in
`codex-rs/core/src/mcp/mod.rs` and `codex-rs/core/src/connectors.rs` to
use the consolidated constructor.
- Removed the now-obsolete `reset_mcp_startup_cancellation_token()`
helper in favor of explicit token replacement at the call sites.
## Testing
- Not run (refactor-only change; no new behavior was intended).
Summary
- expose `agents.max_depth` in config schema and toml parsing, with
defaults and validation
- thread-spawn depth guards and multi-agent handler now respect the
configured limit instead of a hardcoded value
- ensure documentation and helpers account for agent depth limits
TL;DR
Add top-level `model_catalog_json` config support so users can supply a
local model catalog override from a JSON file path (including adding new
models) without backend changes.
### Problem
Codex previously had no clean client-side way to replace/overlay model
catalog data for local testing of model metadata and new model entries.
### Fix
- Add top-level `model_catalog_json` config field (JSON file path).
- Apply catalog entries when resolving `ModelInfo`:
1. Base resolved model metadata (remote/fallback)
2. Catalog overlay from `model_catalog_json`
3. Existing global top-level overrides (`model_context_window`,
`model_supports_reasoning_summaries`, etc.)
### Note
Will revisit per-field overrides in a follow-up
### Tests
Added tests
