This adds `parsed_cmd: Vec<ParsedCommand>` to `ExecApprovalRequestEvent`
in the core protocol (`protocol/src/protocol.rs`), which is also what
this field is named on `ExecCommandBeginEvent`. Honestly, I don't love
the name (it sounds like a single command, but it is actually a list of
them), but I don't want to get distracted by a naming discussion right
now.
This also adds `parsed_cmd` to `ExecCommandApprovalParams` in
`codex-rs/app-server-protocol/src/protocol.rs`, so it will be available
via `codex app-server`, as well.
For consistency, I also updated `ExecApprovalElicitRequestParams` in
`codex-rs/mcp-server/src/exec_approval.rs` to include this field under
the name `codex_parsed_cmd`, as that struct already has a number of
special `codex_*` fields. Note this is the code for when Codex is used
as an MCP _server_ and therefore has to conform to the official spec for
an MCP elicitation type.
We continue the separation between `codex app-server` and `codex
mcp-server`.
In particular, we introduce a new crate, `codex-app-server-protocol`,
and migrate `codex-rs/protocol/src/mcp_protocol.rs` into it, renaming it
`codex-rs/app-server-protocol/src/protocol.rs`.
Because `ConversationId` was defined in `mcp_protocol.rs`, we move it
into its own file, `codex-rs/protocol/src/conversation_id.rs`, and
because it is referenced in a ton of places, we have to touch a lot of
files as part of this PR.
We also decide to get away from proper JSON-RPC 2.0 semantics, so we
also introduce `codex-rs/app-server-protocol/src/jsonrpc_lite.rs`, which
is basically the same `JSONRPCMessage` type defined in `mcp-types`
except with all of the `"jsonrpc": "2.0"` removed.
Getting rid of `"jsonrpc": "2.0"` makes our serialization logic
considerably simpler, as we can lean heavier on serde to serialize
directly into the wire format that we use now.
## 📝 Review Mode -- Core
This PR introduces the Core implementation for Review mode:
- New op `Op::Review { prompt: String }:` spawns a child review task
with isolated context, a review‑specific system prompt, and a
`Config.review_model`.
- `EnteredReviewMode`: emitted when the child review session starts.
Every event from this point onwards reflects the review session.
- `ExitedReviewMode(Option<ReviewOutputEvent>)`: emitted when the review
finishes or is interrupted, with optional structured findings:
```json
{
"findings": [
{
"title": "<≤ 80 chars, imperative>",
"body": "<valid Markdown explaining *why* this is a problem; cite files/lines/functions>",
"confidence_score": <float 0.0-1.0>,
"priority": <int 0-3>,
"code_location": {
"absolute_file_path": "<file path>",
"line_range": {"start": <int>, "end": <int>}
}
}
],
"overall_correctness": "patch is correct" | "patch is incorrect",
"overall_explanation": "<1-3 sentence explanation justifying the overall_correctness verdict>",
"overall_confidence_score": <float 0.0-1.0>
}
```
## Questions
### Why separate out its own message history?
We want the review thread to match the training of our review models as
much as possible -- that means using a custom prompt, removing user
instructions, and starting a clean chat history.
We also want to make sure the review thread doesn't leak into the parent
thread.
### Why do this as a mode, vs. sub-agents?
1. We want review to be a synchronous task, so it's fine for now to do a
bespoke implementation.
2. We're still unclear about the final structure for sub-agents. We'd
prefer to land this quickly and then refactor into sub-agents without
rushing that implementation.
Created this PR by:
- adding `redundant_clone` to `[workspace.lints.clippy]` in
`cargo-rs/Cargol.toml`
- running `cargo clippy --tests --fix`
- running `just fmt`
Though I had to clean up one instance of the following that resulted:
```rust
let codex = codex;
```
This PR changes get history op to get path. Then, forking will use a
path. This will help us have one unified codepath for resuming/forking
conversations. Will also help in having rollout history in order. It
also fixes a bug where you won't see the UI when resuming after forking.
This PR does multiple things that are necessary for conversation resume
to work from the extension. I wanted to make sure everything worked so
these changes wound up in one PR:
1. Generate more ts types
2. Resume rollout history files rather than create a new one every time
it is resumed so you don't see a duplicate conversation in history for
every resume. Chatted with @aibrahim-oai to verify this
3. Return conversation_id in conversation summaries
4. [Cleanup] Use serde and strong types for a lot of the rollout file
parsing
We're trying to migrate from `session_id: Uuid` to `conversation_id:
ConversationId`. Not only does this give us more type safety but it
unifies our terminology across Codex and with the implementation of
session resuming, a conversation (which can span multiple sessions) is
more appropriate.
I started this impl on https://github.com/openai/codex/pull/3219 as part
of getting resume working in the extension but it's big enough that it
should be broken out.
This PR does the following:
- divides user msgs into 3 categories: plain, user instructions, and
environment context
- Centralizes adding user instructions and environment context to a
degree
- Improve the integration testing
Building on top of #3123
Specifically this
[comment](https://github.com/openai/codex/pull/3123#discussion_r2319885089).
We need to send the user message while ignoring the User Instructions
and Environment Context we attach.
- Introduce websearch end to complement the begin
- Moves the logic of adding the sebsearch tool to
create_tools_json_for_responses_api
- Making it the client responsibility to toggle the tool on or off
- Other misc in #2371 post commit feedback
- Show the query:
<img width="1392" height="151" alt="image"
src="https://github.com/user-attachments/assets/8457f1a6-f851-44cf-bcca-0d4fe460ce89"
/>
Adds custom `/prompts` to `~/.codex/prompts/<command>.md`.
<img width="239" height="107" alt="Screenshot 2025-08-25 at 6 22 42 PM"
src="https://github.com/user-attachments/assets/fe6ebbaa-1bf6-49d3-95f9-fdc53b752679"
/>
---
Details:
1. Adds `Op::ListCustomPrompts` to core.
2. Returns `ListCustomPromptsResponse` with list of `CustomPrompt`
(name, content).
3. TUI calls the operation on load, and populates the custom prompts
(excluding prompts that collide with builtins).
4. Selecting the custom prompt automatically sends the prompt to the
agent.
Adds web_search tool, enabling the model to use Responses API web_search
tool.
- Disabled by default, enabled by --search flag
- When --search is passed, exposes web_search_request function tool to
the model, which triggers user approval. When approved, the model can
use the web_search tool for the remainder of the turn
<img width="1033" height="294" alt="image"
src="https://github.com/user-attachments/assets/62ac6563-b946-465c-ba5d-9325af28b28f"
/>
---------
Co-authored-by: easong-openai <easong@openai.com>
This can be the underlying logic in order to start a conversation from a
previous message. will need some love in the UI.
Base for building this: #2588
## Summary
Adds a `/mcp` command to list active tools. We can extend this command
to allow configuration of MCP tools, but for now a simple list command
will help debug if your config.toml and your tools are working as
expected.
Introduces `EventMsg::TurnAborted` that should be sent in response to
`Op::Interrupt`.
In the MCP server, updates the handling of a
`ClientRequest::InterruptConversation` request such that it sends the
`Op::Interrupt` but does not respond to the request until it sees an
`EventMsg::TurnAborted`.
This PR does two things because after I got deep into the first one I
started pulling on the thread to the second:
- Makes `ConversationManager` the place where all in-memory
conversations are created and stored. Previously, `MessageProcessor` in
the `codex-mcp-server` crate was doing this via its `session_map`, but
this is something that should be done in `codex-core`.
- It unwinds the `ctrl_c: tokio::sync::Notify` that was threaded
throughout our code. I think this made sense at one time, but now that
we handle Ctrl-C within the TUI and have a proper `Op::Interrupt` event,
I don't think this was quite right, so I removed it. For `codex exec`
and `codex proto`, we now use `tokio::signal::ctrl_c()` directly, but we
no longer make `Notify` a field of `Codex` or `CodexConversation`.
Changes of note:
- Adds the files `conversation_manager.rs` and `codex_conversation.rs`
to `codex-core`.
- `Codex` and `CodexSpawnOk` are no longer exported from `codex-core`:
other crates must use `CodexConversation` instead (which is created via
`ConversationManager`).
- `core/src/codex_wrapper.rs` has been deleted in favor of
`ConversationManager`.
- `ConversationManager::new_conversation()` returns `NewConversation`,
which is in line with the `new_conversation` tool we want to add to the
MCP server. Note `NewConversation` includes `SessionConfiguredEvent`, so
we eliminate checks in cases like `codex-rs/core/tests/client.rs` to
verify `SessionConfiguredEvent` is the first event because that is now
internal to `ConversationManager`.
- Quite a bit of code was deleted from
`codex-rs/mcp-server/src/message_processor.rs` since it no longer has to
manage multiple conversations itself: it goes through
`ConversationManager` instead.
- `core/tests/live_agent.rs` has been deleted because I had to update a
bunch of tests and all the tests in here were ignored, and I don't think
anyone ever ran them, so this was just technical debt, at this point.
- Removed `notify_on_sigint()` from `util.rs` (and in a follow-up, I
hope to refactor the blandly-named `util.rs` into more descriptive
files).
- In general, I started replacing local variables named `codex` as
`conversation`, where appropriate, though admittedly I didn't do it
through all the integration tests because that would have added a lot of
noise to this PR.
---
[//]: # (BEGIN SAPLING FOOTER)
Stack created with [Sapling](https://sapling-scm.com). Best reviewed
with [ReviewStack](https://reviewstack.dev/openai/codex/pull/2240).
* #2264
* #2263
* __->__ #2240
Wait for newlines, then render markdown on a line by line basis. Word wrap it for the current terminal size and then spit it out line by line into the UI. Also adds tests and fixes some UI regressions.
https://github.com/openai/codex/pull/1835 has some messed up history.
This adds support for streaming chat completions, which is useful for ollama. We should probably take a very skeptical eye to the code introduced in this PR.
---------
Co-authored-by: Ahmed Ibrahim <aibrahim@openai.com>
This lets us show an accumulating diff across all patches in a turn.
Refer to the docs for TurnDiffTracker for implementation details.
There are multiple ways this could have been done and this felt like the
right tradeoff between reliability and completeness:
*Pros*
* It will pick up all changes to files that the model touched including
if they prettier or another command that updates them.
* It will not pick up changes made by the user or other agents to files
it didn't modify.
*Cons*
* It will pick up changes that the user made to a file that the model
also touched
* It will not pick up changes to codegen or files that were not modified
with apply_patch
## Summary
- stream command stdout as `ExecCommandStdout` events
- forward streamed stdout to clients and ignore in human output
processor
- adjust call sites for new streaming API
This adds a tool the model can call to update a plan. The tool doesn't
actually _do_ anything but it gives clients a chance to read and render
the structured plan. We will likely iterate on the prompt and tools
exposed for planning over time.
## Summary
Per the [latest MCP
spec](https://modelcontextprotocol.io/specification/2025-06-18/basic#meta),
the `_meta` field is reserved for metadata. In the [Typescript
Schema](0695a497eb/schema/2025-06-18/schema.ts (L37-L40)),
`progressToken` is defined as a value to be attached to subsequent
notifications for that request.
The
[CallToolRequestParams](0695a497eb/schema/2025-06-18/schema.ts (L806-L817))
extends this definition but overwrites the params field. This ambiguity
makes our generated type definitions tricky, so I'm going to skip
`progressToken` field for now and just send back the `requestId`
instead.
In a future PR, we can clarify, update our `generate_mcp_types.py`
script, and update our progressToken logic accordingly.
## Testing
- [x] Added unit tests
- [x] Manually tested with mcp client
1. Emit call_id to exec approval elicitations for mcp client convenience
2. Remove the `-retry` from the call id for the same reason as above but
upstream the reset behavior to the mcp client
1. Added an elicitation for `approve-patch` which is very similar to
`approve-exec`.
2. Extracted both elicitations to their own files to prevent
`codex_tool_runner` from blowing up in size.
## Summary
Adds a new mcp tool call, `codex-reply`, so we can continue existing
sessions. This is a first draft and does not yet support sessions from
previous processes.
## Testing
- [x] tested with mcp client
This PR introduces a single integration test for `cargo mcp`, though it
also introduces a number of reusable components so that it should be
easier to introduce more integration tests going forward.
The new test is introduced in `codex-rs/mcp-server/tests/elicitation.rs`
and the reusable pieces are in `codex-rs/mcp-server/tests/common`.
The test itself verifies new functionality around elicitations
introduced in https://github.com/openai/codex/pull/1623 (and the fix
introduced in https://github.com/openai/codex/pull/1629) by doing the
following:
- starts a mock model provider with canned responses for
`/v1/chat/completions`
- starts the MCP server with a `config.toml` to use that model provider
(and `approval_policy = "untrusted"`)
- sends the `codex` tool call which causes the mock model provider to
request a shell call for `git init`
- the MCP server sends an elicitation to the client to approve the
request
- the client replies to the elicitation with `"approved"`
- the MCP server runs the command and re-samples the model, getting a
`"finish_reason": "stop"`
- in turn, the MCP server sends the final response to the original
`codex` tool call
- verifies that `git init` ran as expected
To test:
```
cargo test shell_command_approval_triggers_elicitation
```
In writing this test, I discovered that `ExecApprovalResponse` does not
conform to `ElicitResult`, so I added a TODO to fix that, since I think
that should be updated in a separate PR. As it stands, this PR does not
update any business logic, though it does make a number of members of
the `mcp-server` crate `pub` so they can be used in the test.
One additional learning from this PR is that
`std::process::Command::cargo_bin()` from the `assert_cmd` trait is only
available for `std::process::Command`, but we really want to use
`tokio::process::Command` so that everything is async and we can
leverage utilities like `tokio::time::timeout()`. The trick I came up
with was to use `cargo_bin()` to locate the program, and then to use
`std::process::Command::get_program()` when constructing the
`tokio::process::Command`.
This updates the MCP server so that if it receives an
`ExecApprovalRequest` from the `Codex` session, it in turn sends an [MCP
elicitation](https://modelcontextprotocol.io/specification/draft/client/elicitation)
to the client to ask for the approval decision. Upon getting a response,
it forwards the client's decision via `Op::ExecApproval`.
Admittedly, we should be doing the same thing for
`ApplyPatchApprovalRequest`, but this is our first time experimenting
with elicitations, so I'm inclined to defer wiring that code path up
until we feel good about how this one works.
---
[//]: # (BEGIN SAPLING FOOTER)
Stack created with [Sapling](https://sapling-scm.com). Best reviewed
with [ReviewStack](https://reviewstack.dev/openai/codex/pull/1623).
* __->__ #1623
* #1622
* #1621
* #1620
Previous to this change, `MessageProcessor` had a
`tokio::sync::mpsc::Sender<JSONRPCMessage>` as an abstraction for server
code to send a message down to the MCP client. Because `Sender` is cheap
to `clone()`, it was straightforward to make it available to tasks
scheduled with `tokio::task::spawn()`.
This worked well when we were only sending notifications or responses
back down to the client, but we want to add support for sending
elicitations in #1623, which means that we need to be able to send
_requests_ to the client, and now we need a bit of centralization to
ensure all request ids are unique.
To that end, this PR introduces `OutgoingMessageSender`, which houses
the existing `Sender<OutgoingMessage>` as well as an `AtomicI64` to mint
out new, unique request ids. It has methods like `send_request()` and
`send_response()` so that callers do not have to deal with
`JSONRPCMessage` directly, as having to set the `jsonrpc` for each
message was a bit tedious (this cleans up `codex_tool_runner.rs` quite a
bit).
We do not have `OutgoingMessageSender` implement `Clone` because it is
important that the `AtomicI64` is shared across all users of
`OutgoingMessageSender`. As such, `Arc<OutgoingMessageSender>` must be
used instead, as it is frequently shared with new tokio tasks.
As part of this change, we update `message_processor.rs` to embrace
`await`, though we must be careful that no individual handler blocks the
main loop and prevents other messages from being handled.
---
[//]: # (BEGIN SAPLING FOOTER)
Stack created with [Sapling](https://sapling-scm.com). Best reviewed
with [ReviewStack](https://reviewstack.dev/openai/codex/pull/1622).
* #1623
* __->__ #1622
* #1621
* #1620
This updates the schema in `generate_mcp_types.py` from `2025-03-26` to
`2025-06-18`, regenerates `mcp-types/src/lib.rs`, and then updates all
the code that uses `mcp-types` to honor the changes.
Ran
```
npx @modelcontextprotocol/inspector just codex mcp
```
and verified that I was able to invoke the `codex` tool, as expected.
---
[//]: # (BEGIN SAPLING FOOTER)
Stack created with [Sapling](https://sapling-scm.com). Best reviewed
with [ReviewStack](https://reviewstack.dev/openai/codex/pull/1621).
* #1623
* #1622
* __->__ #1621
- Added support for message and reasoning deltas
- Skipped adding the support in the cli and tui for later
- Commented a failing test (wrong merge) that needs fix in a separate
PR.
Side note: I think we need to disable merge when the CI don't pass.
When using the OpenAI Responses API, we now record the `usage` field for
a `"response.completed"` event, which includes metrics about the number
of tokens consumed. We also introduce `openai_model_info.rs`, which
includes current data about the most common OpenAI models available via
the API (specifically `context_window` and `max_output_tokens`). If
Codex does not recognize the model, you can set `model_context_window`
and `model_max_output_tokens` explicitly in `config.toml`.
When then introduce a new event type to `protocol.rs`, `TokenCount`,
which includes the `TokenUsage` for the most recent turn.
Finally, we update the TUI to record the running sum of tokens used so
the percentage of available context window remaining can be reported via
the placeholder text for the composer:
We could certainly get much fancier with this (such as reporting the
estimated cost of the conversation), but for now, we are just trying to
achieve feature parity with the TypeScript CLI.
Though arguably this improves upon the TypeScript CLI, as the TypeScript
CLI uses heuristics to estimate the number of tokens used rather than
using the `usage` information directly:
296996d74e/codex-cli/src/utils/approximate-tokens-used.ts (L3-L16)
Fixes https://github.com/openai/codex/issues/1242
This introduces an experimental `--output-last-message` flag that can be
used to identify a file where the final message from the agent will be
written. Two use cases:
- Ultimately, we will likely add a `--quiet` option to `exec`, but even
if the user does not want any output written to the terminal, they
probably want to know what the agent did. Writing the output to a file
makes it possible to get that information in a clean way.
- Relatedly, when using `exec` in CI, it is easier to review the
transcript written "normally," (i.e., not as JSON or something with
extra escapes), but getting programmatic access to the last message is
likely helpful, so writing the last message to a file gets the best of
both worlds.
I am calling this "experimental" because it is possible that we are
overfitting and will want a more general solution to this problem that
would justify removing this flag.
This is a large change to support a "history" feature like you would
expect in a shell like Bash.
History events are recorded in `$CODEX_HOME/history.jsonl`. Because it
is a JSONL file, it is straightforward to append new entries (as opposed
to the TypeScript file that uses `$CODEX_HOME/history.json`, so to be
valid JSON, each new entry entails rewriting the entire file). Because
it is possible for there to be multiple instances of Codex CLI writing
to `history.jsonl` at once, we use advisory file locking when working
with `history.jsonl` in `codex-rs/core/src/message_history.rs`.
Because we believe history is a sufficiently useful feature, we enable
it by default. Though to provide some safety, we set the file
permissions of `history.jsonl` to be `o600` so that other users on the
system cannot read the user's history. We do not yet support a default
list of `SENSITIVE_PATTERNS` as the TypeScript CLI does:
3fdf9df133/codex-cli/src/utils/storage/command-history.ts (L10-L17)
We are going to take a more conservative approach to this list in the
Rust CLI. For example, while `/\b[A-Za-z0-9-_]{20,}\b/` might exclude
sensitive information like API tokens, it would also exclude valuable
information such as references to Git commits.
As noted in the updated documentation, users can opt-out of history by
adding the following to `config.toml`:
```toml
[history]
persistence = "none"
```
Because `history.jsonl` could, in theory, be quite large, we take a[n
arguably overly pedantic] approach in reading history entries into
memory. Specifically, we start by telling the client the current number
of entries in the history file (`history_entry_count`) as well as the
inode (`history_log_id`) of `history.jsonl` (see the new fields on
`SessionConfiguredEvent`).
The client is responsible for keeping new entries in memory to create a
"local history," but if the user hits up enough times to go "past" the
end of local history, then the client should use the new
`GetHistoryEntryRequest` in the protocol to fetch older entries.
Specifically, it should pass the `history_log_id` it was given
originally and work backwards from `history_entry_count`. (It should
really fetch history in batches rather than one-at-a-time, but that is
something we can improve upon in subsequent PRs.)
The motivation behind this crazy scheme is that it is designed to defend
against:
* The `history.jsonl` being truncated during the session such that the
index into the history is no longer consistent with what had been read
up to that point. We do not yet have logic to enforce a `max_bytes` for
`history.jsonl`, but once we do, we will aspire to implement it in a way
that should result in a new inode for the file on most systems.
* New items from concurrent Codex CLI sessions amending to the history.
Because, in absence of truncation, `history.jsonl` is an append-only
log, so long as the client reads backwards from `history_entry_count`,
it should always get a consistent view of history. (That said, it will
not be able to read _new_ commands from concurrent sessions, but perhaps
we will introduce a `/` command to reload latest history or something
down the road.)
Admittedly, my testing of this feature thus far has been fairly light. I
expect we will find bugs and introduce enhancements/fixes going forward.
For now, this removes the `#[non_exhaustive]` directive on `EventMsg` so
that we are forced to handle all `EventMsg` by default. (We may revisit
this if/when we publish `core/` as a `lib` crate.) For now, it is
helpful to have this as a forcing function because we have effectively
two UIs (`tui` and `exec`) and usually when we add a new variant to
`EventMsg`, we want to be sure that we update both.
https://github.com/openai/codex/pull/922 did this for the
`SessionConfigured` enum variant, and I think it is generally helpful to
be able to work with the values as each enum variant as their own type,
so this converts the remaining variants and updates all of the
callsites.
Added a simple unit test to verify that the JSON-serialized version of
`Event` does not have any unexpected nesting.
