Files
llmx/codex-rs/mcp-server/src/codex_tool_config.rs

336 lines
12 KiB
Rust
Raw Normal View History

feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
//! Configuration object accepted by the `codex` MCP tool-call.
feat: add support for -c/--config to override individual config items (#1137) This PR introduces support for `-c`/`--config` so users can override individual config values on the command line using `--config name=value`. Example: ``` codex --config model=o4-mini ``` Making it possible to set arbitrary config values on the command line results in a more flexible configuration scheme and makes it easier to provide single-line examples that can be copy-pasted from documentation. Effectively, it means there are four levels of configuration for some values: - Default value (e.g., `model` currently defaults to `o4-mini`) - Value in `config.toml` (e.g., user could override the default to be `model = "o3"` in their `config.toml`) - Specifying `-c` or `--config` to override `model` (e.g., user can include `-c model=o3` in their list of args to Codex) - If available, a config-specific flag can be used, which takes precedence over `-c` (e.g., user can specify `--model o3` in their list of args to Codex) Now that it is possible to specify anything that could be configured in `config.toml` on the command line using `-c`, we do not need to have a custom flag for every possible config option (which can clutter the output of `--help`). To that end, as part of this PR, we drop support for the `--disable-response-storage` flag, as users can now specify `-c disable_response_storage=true` to get the equivalent functionality. Under the hood, this works by loading the `config.toml` into a `toml::Value`. Then for each `key=value`, we create a small synthetic TOML file with `value` so that we can run the TOML parser to get the equivalent `toml::Value`. We then parse `key` to determine the point in the original `toml::Value` to do the insert/replace. Once all of the overrides from `-c` args have been applied, the `toml::Value` is deserialized into a `ConfigToml` and then the `ConfigOverrides` are applied, as before.
2025-05-27 23:11:44 -07:00
use codex_core::protocol::AskForApproval;
use codex_protocol::config_types::SandboxMode;
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
use mcp_types::Tool;
use mcp_types::ToolInputSchema;
use schemars::JsonSchema;
use schemars::r#gen::SchemaSettings;
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
use serde::Deserialize;
test: add integration test for MCP server (#1633) 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`.
2025-07-21 10:27:07 -07:00
use serde::Serialize;
feat: add support for -c/--config to override individual config items (#1137) This PR introduces support for `-c`/`--config` so users can override individual config values on the command line using `--config name=value`. Example: ``` codex --config model=o4-mini ``` Making it possible to set arbitrary config values on the command line results in a more flexible configuration scheme and makes it easier to provide single-line examples that can be copy-pasted from documentation. Effectively, it means there are four levels of configuration for some values: - Default value (e.g., `model` currently defaults to `o4-mini`) - Value in `config.toml` (e.g., user could override the default to be `model = "o3"` in their `config.toml`) - Specifying `-c` or `--config` to override `model` (e.g., user can include `-c model=o3` in their list of args to Codex) - If available, a config-specific flag can be used, which takes precedence over `-c` (e.g., user can specify `--model o3` in their list of args to Codex) Now that it is possible to specify anything that could be configured in `config.toml` on the command line using `-c`, we do not need to have a custom flag for every possible config option (which can clutter the output of `--help`). To that end, as part of this PR, we drop support for the `--disable-response-storage` flag, as users can now specify `-c disable_response_storage=true` to get the equivalent functionality. Under the hood, this works by loading the `config.toml` into a `toml::Value`. Then for each `key=value`, we create a small synthetic TOML file with `value` so that we can run the TOML parser to get the equivalent `toml::Value`. We then parse `key` to determine the point in the original `toml::Value` to do the insert/replace. Once all of the overrides from `-c` args have been applied, the `toml::Value` is deserialized into a `ConfigToml` and then the `ConfigOverrides` are applied, as before.
2025-05-27 23:11:44 -07:00
use std::collections::HashMap;
use std::path::PathBuf;
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
feat: add support for -c/--config to override individual config items (#1137) This PR introduces support for `-c`/`--config` so users can override individual config values on the command line using `--config name=value`. Example: ``` codex --config model=o4-mini ``` Making it possible to set arbitrary config values on the command line results in a more flexible configuration scheme and makes it easier to provide single-line examples that can be copy-pasted from documentation. Effectively, it means there are four levels of configuration for some values: - Default value (e.g., `model` currently defaults to `o4-mini`) - Value in `config.toml` (e.g., user could override the default to be `model = "o3"` in their `config.toml`) - Specifying `-c` or `--config` to override `model` (e.g., user can include `-c model=o3` in their list of args to Codex) - If available, a config-specific flag can be used, which takes precedence over `-c` (e.g., user can specify `--model o3` in their list of args to Codex) Now that it is possible to specify anything that could be configured in `config.toml` on the command line using `-c`, we do not need to have a custom flag for every possible config option (which can clutter the output of `--help`). To that end, as part of this PR, we drop support for the `--disable-response-storage` flag, as users can now specify `-c disable_response_storage=true` to get the equivalent functionality. Under the hood, this works by loading the `config.toml` into a `toml::Value`. Then for each `key=value`, we create a small synthetic TOML file with `value` so that we can run the TOML parser to get the equivalent `toml::Value`. We then parse `key` to determine the point in the original `toml::Value` to do the insert/replace. Once all of the overrides from `-c` args have been applied, the `toml::Value` is deserialized into a `ConfigToml` and then the `ConfigOverrides` are applied, as before.
2025-05-27 23:11:44 -07:00
use crate::json_to_toml::json_to_toml;
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
/// Client-supplied configuration for a `codex` tool-call.
#[derive(Debug, Clone, Serialize, Deserialize, JsonSchema, Default)]
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
#[serde(rename_all = "kebab-case")]
test: add integration test for MCP server (#1633) 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`.
2025-07-21 10:27:07 -07:00
pub struct CodexToolCallParam {
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
/// The *initial user prompt* to start the Codex conversation.
pub prompt: String,
/// Optional override for the model name (e.g. "o3", "o4-mini").
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
#[serde(default, skip_serializing_if = "Option::is_none")]
pub model: Option<String>,
/// Configuration profile from config.toml to specify default options.
#[serde(default, skip_serializing_if = "Option::is_none")]
pub profile: Option<String>,
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
/// Working directory for the session. If relative, it is resolved against
/// the server process's current working directory.
#[serde(default, skip_serializing_if = "Option::is_none")]
pub cwd: Option<String>,
feat: add support for --sandbox flag (#1476) On a high-level, we try to design `config.toml` so that you don't have to "comment out a lot of stuff" when testing different options. Previously, defining a sandbox policy was somewhat at odds with this principle because you would define the policy as attributes of `[sandbox]` like so: ```toml [sandbox] mode = "workspace-write" writable_roots = [ "/tmp" ] ``` but if you wanted to temporarily change to a read-only sandbox, you might feel compelled to modify your file to be: ```toml [sandbox] mode = "read-only" # mode = "workspace-write" # writable_roots = [ "/tmp" ] ``` Technically, commenting out `writable_roots` would not be strictly necessary, as `mode = "read-only"` would ignore `writable_roots`, but it's still a reasonable thing to do to keep things tidy. Currently, the various values for `mode` do not support that many attributes, so this is not that hard to maintain, but one could imagine this becoming more complex in the future. In this PR, we change Codex CLI so that it no longer recognizes `[sandbox]`. Instead, it introduces a top-level option, `sandbox_mode`, and `[sandbox_workspace_write]` is used to further configure the sandbox when when `sandbox_mode = "workspace-write"` is used: ```toml sandbox_mode = "workspace-write" [sandbox_workspace_write] writable_roots = [ "/tmp" ] ``` This feels a bit more future-proof in that it is less tedious to configure different sandboxes: ```toml sandbox_mode = "workspace-write" [sandbox_read_only] # read-only options here... [sandbox_workspace_write] writable_roots = [ "/tmp" ] [sandbox_danger_full_access] # danger-full-access options here... ``` In this scheme, you never need to comment out the configuration for an individual sandbox type: you only need to redefine `sandbox_mode`. Relatedly, previous to this change, a user had to do `-c sandbox.mode=read-only` to change the mode on the command line. With this change, things are arguably a bit cleaner because the equivalent option is `-c sandbox_mode=read-only` (and now `-c sandbox_workspace_write=...` can be set separately). Though more importantly, we introduce the `-s/--sandbox` option to the CLI, which maps directly to `sandbox_mode` in `config.toml`, making config override behavior easier to reason about. Moreover, as you can see in the updates to the various Markdown files, it is much easier to explain how to configure sandboxing when things like `--sandbox read-only` can be used as an example. Relatedly, this cleanup also made it straightforward to add support for a `sandbox` option for Codex when used as an MCP server (see the changes to `mcp-server/src/codex_tool_config.rs`). Fixes https://github.com/openai/codex/issues/1248.
2025-07-07 22:31:30 -07:00
/// Approval policy for shell commands generated by the model:
/// `untrusted`, `on-failure`, `on-request`, `never`.
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
#[serde(default, skip_serializing_if = "Option::is_none")]
pub approval_policy: Option<CodexToolCallApprovalPolicy>,
feat: add support for --sandbox flag (#1476) On a high-level, we try to design `config.toml` so that you don't have to "comment out a lot of stuff" when testing different options. Previously, defining a sandbox policy was somewhat at odds with this principle because you would define the policy as attributes of `[sandbox]` like so: ```toml [sandbox] mode = "workspace-write" writable_roots = [ "/tmp" ] ``` but if you wanted to temporarily change to a read-only sandbox, you might feel compelled to modify your file to be: ```toml [sandbox] mode = "read-only" # mode = "workspace-write" # writable_roots = [ "/tmp" ] ``` Technically, commenting out `writable_roots` would not be strictly necessary, as `mode = "read-only"` would ignore `writable_roots`, but it's still a reasonable thing to do to keep things tidy. Currently, the various values for `mode` do not support that many attributes, so this is not that hard to maintain, but one could imagine this becoming more complex in the future. In this PR, we change Codex CLI so that it no longer recognizes `[sandbox]`. Instead, it introduces a top-level option, `sandbox_mode`, and `[sandbox_workspace_write]` is used to further configure the sandbox when when `sandbox_mode = "workspace-write"` is used: ```toml sandbox_mode = "workspace-write" [sandbox_workspace_write] writable_roots = [ "/tmp" ] ``` This feels a bit more future-proof in that it is less tedious to configure different sandboxes: ```toml sandbox_mode = "workspace-write" [sandbox_read_only] # read-only options here... [sandbox_workspace_write] writable_roots = [ "/tmp" ] [sandbox_danger_full_access] # danger-full-access options here... ``` In this scheme, you never need to comment out the configuration for an individual sandbox type: you only need to redefine `sandbox_mode`. Relatedly, previous to this change, a user had to do `-c sandbox.mode=read-only` to change the mode on the command line. With this change, things are arguably a bit cleaner because the equivalent option is `-c sandbox_mode=read-only` (and now `-c sandbox_workspace_write=...` can be set separately). Though more importantly, we introduce the `-s/--sandbox` option to the CLI, which maps directly to `sandbox_mode` in `config.toml`, making config override behavior easier to reason about. Moreover, as you can see in the updates to the various Markdown files, it is much easier to explain how to configure sandboxing when things like `--sandbox read-only` can be used as an example. Relatedly, this cleanup also made it straightforward to add support for a `sandbox` option for Codex when used as an MCP server (see the changes to `mcp-server/src/codex_tool_config.rs`). Fixes https://github.com/openai/codex/issues/1248.
2025-07-07 22:31:30 -07:00
/// Sandbox mode: `read-only`, `workspace-write`, or `danger-full-access`.
#[serde(default, skip_serializing_if = "Option::is_none")]
pub sandbox: Option<CodexToolCallSandboxMode>,
feat: add support for -c/--config to override individual config items (#1137) This PR introduces support for `-c`/`--config` so users can override individual config values on the command line using `--config name=value`. Example: ``` codex --config model=o4-mini ``` Making it possible to set arbitrary config values on the command line results in a more flexible configuration scheme and makes it easier to provide single-line examples that can be copy-pasted from documentation. Effectively, it means there are four levels of configuration for some values: - Default value (e.g., `model` currently defaults to `o4-mini`) - Value in `config.toml` (e.g., user could override the default to be `model = "o3"` in their `config.toml`) - Specifying `-c` or `--config` to override `model` (e.g., user can include `-c model=o3` in their list of args to Codex) - If available, a config-specific flag can be used, which takes precedence over `-c` (e.g., user can specify `--model o3` in their list of args to Codex) Now that it is possible to specify anything that could be configured in `config.toml` on the command line using `-c`, we do not need to have a custom flag for every possible config option (which can clutter the output of `--help`). To that end, as part of this PR, we drop support for the `--disable-response-storage` flag, as users can now specify `-c disable_response_storage=true` to get the equivalent functionality. Under the hood, this works by loading the `config.toml` into a `toml::Value`. Then for each `key=value`, we create a small synthetic TOML file with `value` so that we can run the TOML parser to get the equivalent `toml::Value`. We then parse `key` to determine the point in the original `toml::Value` to do the insert/replace. Once all of the overrides from `-c` args have been applied, the `toml::Value` is deserialized into a `ConfigToml` and then the `ConfigOverrides` are applied, as before.
2025-05-27 23:11:44 -07:00
/// Individual config settings that will override what is in
/// CODEX_HOME/config.toml.
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
#[serde(default, skip_serializing_if = "Option::is_none")]
feat: add support for -c/--config to override individual config items (#1137) This PR introduces support for `-c`/`--config` so users can override individual config values on the command line using `--config name=value`. Example: ``` codex --config model=o4-mini ``` Making it possible to set arbitrary config values on the command line results in a more flexible configuration scheme and makes it easier to provide single-line examples that can be copy-pasted from documentation. Effectively, it means there are four levels of configuration for some values: - Default value (e.g., `model` currently defaults to `o4-mini`) - Value in `config.toml` (e.g., user could override the default to be `model = "o3"` in their `config.toml`) - Specifying `-c` or `--config` to override `model` (e.g., user can include `-c model=o3` in their list of args to Codex) - If available, a config-specific flag can be used, which takes precedence over `-c` (e.g., user can specify `--model o3` in their list of args to Codex) Now that it is possible to specify anything that could be configured in `config.toml` on the command line using `-c`, we do not need to have a custom flag for every possible config option (which can clutter the output of `--help`). To that end, as part of this PR, we drop support for the `--disable-response-storage` flag, as users can now specify `-c disable_response_storage=true` to get the equivalent functionality. Under the hood, this works by loading the `config.toml` into a `toml::Value`. Then for each `key=value`, we create a small synthetic TOML file with `value` so that we can run the TOML parser to get the equivalent `toml::Value`. We then parse `key` to determine the point in the original `toml::Value` to do the insert/replace. Once all of the overrides from `-c` args have been applied, the `toml::Value` is deserialized into a `ConfigToml` and then the `ConfigOverrides` are applied, as before.
2025-05-27 23:11:44 -07:00
pub config: Option<HashMap<String, serde_json::Value>>,
/// The set of instructions to use instead of the default ones.
#[serde(default, skip_serializing_if = "Option::is_none")]
pub base_instructions: Option<String>,
/// Whether to include the plan tool in the conversation.
#[serde(default, skip_serializing_if = "Option::is_none")]
pub include_plan_tool: Option<bool>,
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
}
feat: add support for --sandbox flag (#1476) On a high-level, we try to design `config.toml` so that you don't have to "comment out a lot of stuff" when testing different options. Previously, defining a sandbox policy was somewhat at odds with this principle because you would define the policy as attributes of `[sandbox]` like so: ```toml [sandbox] mode = "workspace-write" writable_roots = [ "/tmp" ] ``` but if you wanted to temporarily change to a read-only sandbox, you might feel compelled to modify your file to be: ```toml [sandbox] mode = "read-only" # mode = "workspace-write" # writable_roots = [ "/tmp" ] ``` Technically, commenting out `writable_roots` would not be strictly necessary, as `mode = "read-only"` would ignore `writable_roots`, but it's still a reasonable thing to do to keep things tidy. Currently, the various values for `mode` do not support that many attributes, so this is not that hard to maintain, but one could imagine this becoming more complex in the future. In this PR, we change Codex CLI so that it no longer recognizes `[sandbox]`. Instead, it introduces a top-level option, `sandbox_mode`, and `[sandbox_workspace_write]` is used to further configure the sandbox when when `sandbox_mode = "workspace-write"` is used: ```toml sandbox_mode = "workspace-write" [sandbox_workspace_write] writable_roots = [ "/tmp" ] ``` This feels a bit more future-proof in that it is less tedious to configure different sandboxes: ```toml sandbox_mode = "workspace-write" [sandbox_read_only] # read-only options here... [sandbox_workspace_write] writable_roots = [ "/tmp" ] [sandbox_danger_full_access] # danger-full-access options here... ``` In this scheme, you never need to comment out the configuration for an individual sandbox type: you only need to redefine `sandbox_mode`. Relatedly, previous to this change, a user had to do `-c sandbox.mode=read-only` to change the mode on the command line. With this change, things are arguably a bit cleaner because the equivalent option is `-c sandbox_mode=read-only` (and now `-c sandbox_workspace_write=...` can be set separately). Though more importantly, we introduce the `-s/--sandbox` option to the CLI, which maps directly to `sandbox_mode` in `config.toml`, making config override behavior easier to reason about. Moreover, as you can see in the updates to the various Markdown files, it is much easier to explain how to configure sandboxing when things like `--sandbox read-only` can be used as an example. Relatedly, this cleanup also made it straightforward to add support for a `sandbox` option for Codex when used as an MCP server (see the changes to `mcp-server/src/codex_tool_config.rs`). Fixes https://github.com/openai/codex/issues/1248.
2025-07-07 22:31:30 -07:00
/// Custom enum mirroring [`AskForApproval`], but has an extra dependency on
/// [`JsonSchema`].
feat: support traditional JSON-RPC request/response in MCP server (#2264) This introduces a new set of request types that our `codex mcp` supports. Note that these do not conform to MCP tool calls so that instead of having to send something like this: ```json { "jsonrpc": "2.0", "method": "tools/call", "id": 42, "params": { "name": "newConversation", "arguments": { "model": "gpt-5", "approvalPolicy": "on-request" } } } ``` we can send something like this: ```json { "jsonrpc": "2.0", "method": "newConversation", "id": 42, "params": { "model": "gpt-5", "approvalPolicy": "on-request" } } ``` Admittedly, this new format is not a valid MCP tool call, but we are OK with that right now. (That is, not everything we might want to request of `codex mcp` is something that is appropriate for an autonomous agent to do.) To start, this introduces four request types: - `newConversation` - `sendUserMessage` - `addConversationListener` - `removeConversationListener` The new `mcp-server/tests/codex_message_processor_flow.rs` shows how these can be used. The types are defined on the `CodexRequest` enum, so we introduce a new `CodexMessageProcessor` that is responsible for dealing with requests from this enum. The top-level `MessageProcessor` has been updated so that when `process_request()` is called, it first checks whether the request conforms to `CodexRequest` and dispatches it to `CodexMessageProcessor` if so. Note that I also decided to use `camelCase` for the on-the-wire format, as that seems to be the convention for MCP. For the moment, the new protocol is defined in `wire_format.rs` within the `mcp-server` crate, but in a subsequent PR, I will probably move it to its own crate to ensure the protocol has minimal dependencies and that we can codegen a schema from it. --- [//]: # (BEGIN SAPLING FOOTER) Stack created with [Sapling](https://sapling-scm.com). Best reviewed with [ReviewStack](https://reviewstack.dev/openai/codex/pull/2264). * #2278 * __->__ #2264
2025-08-13 17:36:29 -07:00
#[derive(Debug, Clone, Serialize, Deserialize, JsonSchema, PartialEq)]
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
#[serde(rename_all = "kebab-case")]
test: add integration test for MCP server (#1633) 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`.
2025-07-21 10:27:07 -07:00
pub enum CodexToolCallApprovalPolicy {
Untrusted,
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
OnFailure,
OnRequest,
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
Never,
}
impl From<CodexToolCallApprovalPolicy> for AskForApproval {
fn from(value: CodexToolCallApprovalPolicy) -> Self {
match value {
CodexToolCallApprovalPolicy::Untrusted => AskForApproval::UnlessTrusted,
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
CodexToolCallApprovalPolicy::OnFailure => AskForApproval::OnFailure,
CodexToolCallApprovalPolicy::OnRequest => AskForApproval::OnRequest,
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
CodexToolCallApprovalPolicy::Never => AskForApproval::Never,
}
}
}
feat: add support for --sandbox flag (#1476) On a high-level, we try to design `config.toml` so that you don't have to "comment out a lot of stuff" when testing different options. Previously, defining a sandbox policy was somewhat at odds with this principle because you would define the policy as attributes of `[sandbox]` like so: ```toml [sandbox] mode = "workspace-write" writable_roots = [ "/tmp" ] ``` but if you wanted to temporarily change to a read-only sandbox, you might feel compelled to modify your file to be: ```toml [sandbox] mode = "read-only" # mode = "workspace-write" # writable_roots = [ "/tmp" ] ``` Technically, commenting out `writable_roots` would not be strictly necessary, as `mode = "read-only"` would ignore `writable_roots`, but it's still a reasonable thing to do to keep things tidy. Currently, the various values for `mode` do not support that many attributes, so this is not that hard to maintain, but one could imagine this becoming more complex in the future. In this PR, we change Codex CLI so that it no longer recognizes `[sandbox]`. Instead, it introduces a top-level option, `sandbox_mode`, and `[sandbox_workspace_write]` is used to further configure the sandbox when when `sandbox_mode = "workspace-write"` is used: ```toml sandbox_mode = "workspace-write" [sandbox_workspace_write] writable_roots = [ "/tmp" ] ``` This feels a bit more future-proof in that it is less tedious to configure different sandboxes: ```toml sandbox_mode = "workspace-write" [sandbox_read_only] # read-only options here... [sandbox_workspace_write] writable_roots = [ "/tmp" ] [sandbox_danger_full_access] # danger-full-access options here... ``` In this scheme, you never need to comment out the configuration for an individual sandbox type: you only need to redefine `sandbox_mode`. Relatedly, previous to this change, a user had to do `-c sandbox.mode=read-only` to change the mode on the command line. With this change, things are arguably a bit cleaner because the equivalent option is `-c sandbox_mode=read-only` (and now `-c sandbox_workspace_write=...` can be set separately). Though more importantly, we introduce the `-s/--sandbox` option to the CLI, which maps directly to `sandbox_mode` in `config.toml`, making config override behavior easier to reason about. Moreover, as you can see in the updates to the various Markdown files, it is much easier to explain how to configure sandboxing when things like `--sandbox read-only` can be used as an example. Relatedly, this cleanup also made it straightforward to add support for a `sandbox` option for Codex when used as an MCP server (see the changes to `mcp-server/src/codex_tool_config.rs`). Fixes https://github.com/openai/codex/issues/1248.
2025-07-07 22:31:30 -07:00
/// Custom enum mirroring [`SandboxMode`] from config_types.rs, but with
/// `JsonSchema` support.
feat: support traditional JSON-RPC request/response in MCP server (#2264) This introduces a new set of request types that our `codex mcp` supports. Note that these do not conform to MCP tool calls so that instead of having to send something like this: ```json { "jsonrpc": "2.0", "method": "tools/call", "id": 42, "params": { "name": "newConversation", "arguments": { "model": "gpt-5", "approvalPolicy": "on-request" } } } ``` we can send something like this: ```json { "jsonrpc": "2.0", "method": "newConversation", "id": 42, "params": { "model": "gpt-5", "approvalPolicy": "on-request" } } ``` Admittedly, this new format is not a valid MCP tool call, but we are OK with that right now. (That is, not everything we might want to request of `codex mcp` is something that is appropriate for an autonomous agent to do.) To start, this introduces four request types: - `newConversation` - `sendUserMessage` - `addConversationListener` - `removeConversationListener` The new `mcp-server/tests/codex_message_processor_flow.rs` shows how these can be used. The types are defined on the `CodexRequest` enum, so we introduce a new `CodexMessageProcessor` that is responsible for dealing with requests from this enum. The top-level `MessageProcessor` has been updated so that when `process_request()` is called, it first checks whether the request conforms to `CodexRequest` and dispatches it to `CodexMessageProcessor` if so. Note that I also decided to use `camelCase` for the on-the-wire format, as that seems to be the convention for MCP. For the moment, the new protocol is defined in `wire_format.rs` within the `mcp-server` crate, but in a subsequent PR, I will probably move it to its own crate to ensure the protocol has minimal dependencies and that we can codegen a schema from it. --- [//]: # (BEGIN SAPLING FOOTER) Stack created with [Sapling](https://sapling-scm.com). Best reviewed with [ReviewStack](https://reviewstack.dev/openai/codex/pull/2264). * #2278 * __->__ #2264
2025-08-13 17:36:29 -07:00
#[derive(Debug, Clone, Serialize, Deserialize, JsonSchema, PartialEq)]
feat: add support for --sandbox flag (#1476) On a high-level, we try to design `config.toml` so that you don't have to "comment out a lot of stuff" when testing different options. Previously, defining a sandbox policy was somewhat at odds with this principle because you would define the policy as attributes of `[sandbox]` like so: ```toml [sandbox] mode = "workspace-write" writable_roots = [ "/tmp" ] ``` but if you wanted to temporarily change to a read-only sandbox, you might feel compelled to modify your file to be: ```toml [sandbox] mode = "read-only" # mode = "workspace-write" # writable_roots = [ "/tmp" ] ``` Technically, commenting out `writable_roots` would not be strictly necessary, as `mode = "read-only"` would ignore `writable_roots`, but it's still a reasonable thing to do to keep things tidy. Currently, the various values for `mode` do not support that many attributes, so this is not that hard to maintain, but one could imagine this becoming more complex in the future. In this PR, we change Codex CLI so that it no longer recognizes `[sandbox]`. Instead, it introduces a top-level option, `sandbox_mode`, and `[sandbox_workspace_write]` is used to further configure the sandbox when when `sandbox_mode = "workspace-write"` is used: ```toml sandbox_mode = "workspace-write" [sandbox_workspace_write] writable_roots = [ "/tmp" ] ``` This feels a bit more future-proof in that it is less tedious to configure different sandboxes: ```toml sandbox_mode = "workspace-write" [sandbox_read_only] # read-only options here... [sandbox_workspace_write] writable_roots = [ "/tmp" ] [sandbox_danger_full_access] # danger-full-access options here... ``` In this scheme, you never need to comment out the configuration for an individual sandbox type: you only need to redefine `sandbox_mode`. Relatedly, previous to this change, a user had to do `-c sandbox.mode=read-only` to change the mode on the command line. With this change, things are arguably a bit cleaner because the equivalent option is `-c sandbox_mode=read-only` (and now `-c sandbox_workspace_write=...` can be set separately). Though more importantly, we introduce the `-s/--sandbox` option to the CLI, which maps directly to `sandbox_mode` in `config.toml`, making config override behavior easier to reason about. Moreover, as you can see in the updates to the various Markdown files, it is much easier to explain how to configure sandboxing when things like `--sandbox read-only` can be used as an example. Relatedly, this cleanup also made it straightforward to add support for a `sandbox` option for Codex when used as an MCP server (see the changes to `mcp-server/src/codex_tool_config.rs`). Fixes https://github.com/openai/codex/issues/1248.
2025-07-07 22:31:30 -07:00
#[serde(rename_all = "kebab-case")]
test: add integration test for MCP server (#1633) 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`.
2025-07-21 10:27:07 -07:00
pub enum CodexToolCallSandboxMode {
feat: add support for --sandbox flag (#1476) On a high-level, we try to design `config.toml` so that you don't have to "comment out a lot of stuff" when testing different options. Previously, defining a sandbox policy was somewhat at odds with this principle because you would define the policy as attributes of `[sandbox]` like so: ```toml [sandbox] mode = "workspace-write" writable_roots = [ "/tmp" ] ``` but if you wanted to temporarily change to a read-only sandbox, you might feel compelled to modify your file to be: ```toml [sandbox] mode = "read-only" # mode = "workspace-write" # writable_roots = [ "/tmp" ] ``` Technically, commenting out `writable_roots` would not be strictly necessary, as `mode = "read-only"` would ignore `writable_roots`, but it's still a reasonable thing to do to keep things tidy. Currently, the various values for `mode` do not support that many attributes, so this is not that hard to maintain, but one could imagine this becoming more complex in the future. In this PR, we change Codex CLI so that it no longer recognizes `[sandbox]`. Instead, it introduces a top-level option, `sandbox_mode`, and `[sandbox_workspace_write]` is used to further configure the sandbox when when `sandbox_mode = "workspace-write"` is used: ```toml sandbox_mode = "workspace-write" [sandbox_workspace_write] writable_roots = [ "/tmp" ] ``` This feels a bit more future-proof in that it is less tedious to configure different sandboxes: ```toml sandbox_mode = "workspace-write" [sandbox_read_only] # read-only options here... [sandbox_workspace_write] writable_roots = [ "/tmp" ] [sandbox_danger_full_access] # danger-full-access options here... ``` In this scheme, you never need to comment out the configuration for an individual sandbox type: you only need to redefine `sandbox_mode`. Relatedly, previous to this change, a user had to do `-c sandbox.mode=read-only` to change the mode on the command line. With this change, things are arguably a bit cleaner because the equivalent option is `-c sandbox_mode=read-only` (and now `-c sandbox_workspace_write=...` can be set separately). Though more importantly, we introduce the `-s/--sandbox` option to the CLI, which maps directly to `sandbox_mode` in `config.toml`, making config override behavior easier to reason about. Moreover, as you can see in the updates to the various Markdown files, it is much easier to explain how to configure sandboxing when things like `--sandbox read-only` can be used as an example. Relatedly, this cleanup also made it straightforward to add support for a `sandbox` option for Codex when used as an MCP server (see the changes to `mcp-server/src/codex_tool_config.rs`). Fixes https://github.com/openai/codex/issues/1248.
2025-07-07 22:31:30 -07:00
ReadOnly,
WorkspaceWrite,
DangerFullAccess,
}
impl From<CodexToolCallSandboxMode> for SandboxMode {
fn from(value: CodexToolCallSandboxMode) -> Self {
match value {
CodexToolCallSandboxMode::ReadOnly => SandboxMode::ReadOnly,
CodexToolCallSandboxMode::WorkspaceWrite => SandboxMode::WorkspaceWrite,
CodexToolCallSandboxMode::DangerFullAccess => SandboxMode::DangerFullAccess,
}
}
}
/// Builds a `Tool` definition (JSON schema etc.) for the Codex tool-call.
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
pub(crate) fn create_tool_for_codex_tool_call_param() -> Tool {
let schema = SchemaSettings::draft2019_09()
.with(|s| {
s.inline_subschemas = true;
s.option_add_null_type = false;
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
})
.into_generator()
.into_root_schema_for::<CodexToolCallParam>();
#[expect(clippy::expect_used)]
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
let schema_value =
serde_json::to_value(&schema).expect("Codex tool schema should serialise to JSON");
let tool_input_schema =
serde_json::from_value::<ToolInputSchema>(schema_value).unwrap_or_else(|e| {
panic!("failed to create Tool from schema: {e}");
});
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
Tool {
name: "codex".to_string(),
title: Some("Codex".to_string()),
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
input_schema: tool_input_schema,
// TODO(mbolin): This should be defined.
output_schema: None,
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
description: Some(
"Run a Codex session. Accepts configuration parameters matching the Codex Config struct.".to_string(),
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
),
annotations: None,
}
}
impl CodexToolCallParam {
/// Returns the initial user prompt to start the Codex conversation and the
/// effective Config object generated from the supplied parameters.
pub fn into_config(
self,
codex_linux_sandbox_exe: Option<PathBuf>,
) -> std::io::Result<(String, codex_core::config::Config)> {
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
let Self {
prompt,
model,
profile,
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
cwd,
approval_policy,
feat: add support for --sandbox flag (#1476) On a high-level, we try to design `config.toml` so that you don't have to "comment out a lot of stuff" when testing different options. Previously, defining a sandbox policy was somewhat at odds with this principle because you would define the policy as attributes of `[sandbox]` like so: ```toml [sandbox] mode = "workspace-write" writable_roots = [ "/tmp" ] ``` but if you wanted to temporarily change to a read-only sandbox, you might feel compelled to modify your file to be: ```toml [sandbox] mode = "read-only" # mode = "workspace-write" # writable_roots = [ "/tmp" ] ``` Technically, commenting out `writable_roots` would not be strictly necessary, as `mode = "read-only"` would ignore `writable_roots`, but it's still a reasonable thing to do to keep things tidy. Currently, the various values for `mode` do not support that many attributes, so this is not that hard to maintain, but one could imagine this becoming more complex in the future. In this PR, we change Codex CLI so that it no longer recognizes `[sandbox]`. Instead, it introduces a top-level option, `sandbox_mode`, and `[sandbox_workspace_write]` is used to further configure the sandbox when when `sandbox_mode = "workspace-write"` is used: ```toml sandbox_mode = "workspace-write" [sandbox_workspace_write] writable_roots = [ "/tmp" ] ``` This feels a bit more future-proof in that it is less tedious to configure different sandboxes: ```toml sandbox_mode = "workspace-write" [sandbox_read_only] # read-only options here... [sandbox_workspace_write] writable_roots = [ "/tmp" ] [sandbox_danger_full_access] # danger-full-access options here... ``` In this scheme, you never need to comment out the configuration for an individual sandbox type: you only need to redefine `sandbox_mode`. Relatedly, previous to this change, a user had to do `-c sandbox.mode=read-only` to change the mode on the command line. With this change, things are arguably a bit cleaner because the equivalent option is `-c sandbox_mode=read-only` (and now `-c sandbox_workspace_write=...` can be set separately). Though more importantly, we introduce the `-s/--sandbox` option to the CLI, which maps directly to `sandbox_mode` in `config.toml`, making config override behavior easier to reason about. Moreover, as you can see in the updates to the various Markdown files, it is much easier to explain how to configure sandboxing when things like `--sandbox read-only` can be used as an example. Relatedly, this cleanup also made it straightforward to add support for a `sandbox` option for Codex when used as an MCP server (see the changes to `mcp-server/src/codex_tool_config.rs`). Fixes https://github.com/openai/codex/issues/1248.
2025-07-07 22:31:30 -07:00
sandbox,
feat: add support for -c/--config to override individual config items (#1137) This PR introduces support for `-c`/`--config` so users can override individual config values on the command line using `--config name=value`. Example: ``` codex --config model=o4-mini ``` Making it possible to set arbitrary config values on the command line results in a more flexible configuration scheme and makes it easier to provide single-line examples that can be copy-pasted from documentation. Effectively, it means there are four levels of configuration for some values: - Default value (e.g., `model` currently defaults to `o4-mini`) - Value in `config.toml` (e.g., user could override the default to be `model = "o3"` in their `config.toml`) - Specifying `-c` or `--config` to override `model` (e.g., user can include `-c model=o3` in their list of args to Codex) - If available, a config-specific flag can be used, which takes precedence over `-c` (e.g., user can specify `--model o3` in their list of args to Codex) Now that it is possible to specify anything that could be configured in `config.toml` on the command line using `-c`, we do not need to have a custom flag for every possible config option (which can clutter the output of `--help`). To that end, as part of this PR, we drop support for the `--disable-response-storage` flag, as users can now specify `-c disable_response_storage=true` to get the equivalent functionality. Under the hood, this works by loading the `config.toml` into a `toml::Value`. Then for each `key=value`, we create a small synthetic TOML file with `value` so that we can run the TOML parser to get the equivalent `toml::Value`. We then parse `key` to determine the point in the original `toml::Value` to do the insert/replace. Once all of the overrides from `-c` args have been applied, the `toml::Value` is deserialized into a `ConfigToml` and then the `ConfigOverrides` are applied, as before.
2025-05-27 23:11:44 -07:00
config: cli_overrides,
base_instructions,
include_plan_tool,
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
} = self;
// Build the `ConfigOverrides` recognized by codex-core.
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
let overrides = codex_core::config::ConfigOverrides {
model,
config_profile: profile,
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
cwd: cwd.map(PathBuf::from),
approval_policy: approval_policy.map(Into::into),
feat: add support for --sandbox flag (#1476) On a high-level, we try to design `config.toml` so that you don't have to "comment out a lot of stuff" when testing different options. Previously, defining a sandbox policy was somewhat at odds with this principle because you would define the policy as attributes of `[sandbox]` like so: ```toml [sandbox] mode = "workspace-write" writable_roots = [ "/tmp" ] ``` but if you wanted to temporarily change to a read-only sandbox, you might feel compelled to modify your file to be: ```toml [sandbox] mode = "read-only" # mode = "workspace-write" # writable_roots = [ "/tmp" ] ``` Technically, commenting out `writable_roots` would not be strictly necessary, as `mode = "read-only"` would ignore `writable_roots`, but it's still a reasonable thing to do to keep things tidy. Currently, the various values for `mode` do not support that many attributes, so this is not that hard to maintain, but one could imagine this becoming more complex in the future. In this PR, we change Codex CLI so that it no longer recognizes `[sandbox]`. Instead, it introduces a top-level option, `sandbox_mode`, and `[sandbox_workspace_write]` is used to further configure the sandbox when when `sandbox_mode = "workspace-write"` is used: ```toml sandbox_mode = "workspace-write" [sandbox_workspace_write] writable_roots = [ "/tmp" ] ``` This feels a bit more future-proof in that it is less tedious to configure different sandboxes: ```toml sandbox_mode = "workspace-write" [sandbox_read_only] # read-only options here... [sandbox_workspace_write] writable_roots = [ "/tmp" ] [sandbox_danger_full_access] # danger-full-access options here... ``` In this scheme, you never need to comment out the configuration for an individual sandbox type: you only need to redefine `sandbox_mode`. Relatedly, previous to this change, a user had to do `-c sandbox.mode=read-only` to change the mode on the command line. With this change, things are arguably a bit cleaner because the equivalent option is `-c sandbox_mode=read-only` (and now `-c sandbox_workspace_write=...` can be set separately). Though more importantly, we introduce the `-s/--sandbox` option to the CLI, which maps directly to `sandbox_mode` in `config.toml`, making config override behavior easier to reason about. Moreover, as you can see in the updates to the various Markdown files, it is much easier to explain how to configure sandboxing when things like `--sandbox read-only` can be used as an example. Relatedly, this cleanup also made it straightforward to add support for a `sandbox` option for Codex when used as an MCP server (see the changes to `mcp-server/src/codex_tool_config.rs`). Fixes https://github.com/openai/codex/issues/1248.
2025-07-07 22:31:30 -07:00
sandbox_mode: sandbox.map(Into::into),
model_provider: None,
codex_linux_sandbox_exe,
base_instructions,
include_plan_tool,
include_apply_patch_tool: None,
include_view_image_tool: None,
show_raw_agent_reasoning: None,
tools_web_search_request: None,
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
};
feat: add support for -c/--config to override individual config items (#1137) This PR introduces support for `-c`/`--config` so users can override individual config values on the command line using `--config name=value`. Example: ``` codex --config model=o4-mini ``` Making it possible to set arbitrary config values on the command line results in a more flexible configuration scheme and makes it easier to provide single-line examples that can be copy-pasted from documentation. Effectively, it means there are four levels of configuration for some values: - Default value (e.g., `model` currently defaults to `o4-mini`) - Value in `config.toml` (e.g., user could override the default to be `model = "o3"` in their `config.toml`) - Specifying `-c` or `--config` to override `model` (e.g., user can include `-c model=o3` in their list of args to Codex) - If available, a config-specific flag can be used, which takes precedence over `-c` (e.g., user can specify `--model o3` in their list of args to Codex) Now that it is possible to specify anything that could be configured in `config.toml` on the command line using `-c`, we do not need to have a custom flag for every possible config option (which can clutter the output of `--help`). To that end, as part of this PR, we drop support for the `--disable-response-storage` flag, as users can now specify `-c disable_response_storage=true` to get the equivalent functionality. Under the hood, this works by loading the `config.toml` into a `toml::Value`. Then for each `key=value`, we create a small synthetic TOML file with `value` so that we can run the TOML parser to get the equivalent `toml::Value`. We then parse `key` to determine the point in the original `toml::Value` to do the insert/replace. Once all of the overrides from `-c` args have been applied, the `toml::Value` is deserialized into a `ConfigToml` and then the `ConfigOverrides` are applied, as before.
2025-05-27 23:11:44 -07:00
let cli_overrides = cli_overrides
.unwrap_or_default()
.into_iter()
.map(|(k, v)| (k, json_to_toml(v)))
.collect();
let cfg = codex_core::config::Config::load_with_cli_overrides(cli_overrides, overrides)?;
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
Ok((prompt, cfg))
}
}
#[derive(Debug, Clone, Serialize, Deserialize, JsonSchema)]
#[serde(rename_all = "camelCase")]
pub struct CodexToolCallReplyParam {
/// The conversation id for this Codex session.
pub conversation_id: String,
/// The *next user prompt* to continue the Codex conversation.
pub prompt: String,
}
/// Builds a `Tool` definition for the `codex-reply` tool-call.
pub(crate) fn create_tool_for_codex_tool_call_reply_param() -> Tool {
let schema = SchemaSettings::draft2019_09()
.with(|s| {
s.inline_subschemas = true;
s.option_add_null_type = false;
})
.into_generator()
.into_root_schema_for::<CodexToolCallReplyParam>();
#[expect(clippy::expect_used)]
let schema_value =
serde_json::to_value(&schema).expect("Codex reply tool schema should serialise to JSON");
let tool_input_schema =
serde_json::from_value::<ToolInputSchema>(schema_value).unwrap_or_else(|e| {
panic!("failed to create Tool from schema: {e}");
});
Tool {
name: "codex-reply".to_string(),
title: Some("Codex Reply".to_string()),
input_schema: tool_input_schema,
output_schema: None,
description: Some(
"Continue a Codex conversation by providing the conversation id and prompt."
.to_string(),
),
annotations: None,
}
}
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
#[cfg(test)]
mod tests {
use super::*;
use pretty_assertions::assert_eq;
/// We include a test to verify the exact JSON schema as "executable
/// documentation" for the schema. When can track changes to this test as a
/// way to audit changes to the generated schema.
///
/// Seeing the fully expanded schema makes it easier to casually verify that
/// the generated JSON for enum types such as "approval-policy" is compact.
/// Ideally, modelcontextprotocol/inspector would provide a simpler UI for
/// enum fields versus open string fields to take advantage of this.
///
/// As of 2025-05-04, there is an open PR for this:
/// https://github.com/modelcontextprotocol/inspector/pull/196
#[test]
fn verify_codex_tool_json_schema() {
let tool = create_tool_for_codex_tool_call_param();
let tool_json = serde_json::to_value(&tool).expect("tool serializes");
let expected_tool_json = serde_json::json!({
"name": "codex",
"title": "Codex",
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
"description": "Run a Codex session. Accepts configuration parameters matching the Codex Config struct.",
"inputSchema": {
"type": "object",
"properties": {
"approval-policy": {
"description": "Approval policy for shell commands generated by the model: `untrusted`, `on-failure`, `on-request`, `never`.",
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
"enum": [
"untrusted",
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
"on-failure",
"on-request",
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
"never"
],
"type": "string"
},
feat: add support for --sandbox flag (#1476) On a high-level, we try to design `config.toml` so that you don't have to "comment out a lot of stuff" when testing different options. Previously, defining a sandbox policy was somewhat at odds with this principle because you would define the policy as attributes of `[sandbox]` like so: ```toml [sandbox] mode = "workspace-write" writable_roots = [ "/tmp" ] ``` but if you wanted to temporarily change to a read-only sandbox, you might feel compelled to modify your file to be: ```toml [sandbox] mode = "read-only" # mode = "workspace-write" # writable_roots = [ "/tmp" ] ``` Technically, commenting out `writable_roots` would not be strictly necessary, as `mode = "read-only"` would ignore `writable_roots`, but it's still a reasonable thing to do to keep things tidy. Currently, the various values for `mode` do not support that many attributes, so this is not that hard to maintain, but one could imagine this becoming more complex in the future. In this PR, we change Codex CLI so that it no longer recognizes `[sandbox]`. Instead, it introduces a top-level option, `sandbox_mode`, and `[sandbox_workspace_write]` is used to further configure the sandbox when when `sandbox_mode = "workspace-write"` is used: ```toml sandbox_mode = "workspace-write" [sandbox_workspace_write] writable_roots = [ "/tmp" ] ``` This feels a bit more future-proof in that it is less tedious to configure different sandboxes: ```toml sandbox_mode = "workspace-write" [sandbox_read_only] # read-only options here... [sandbox_workspace_write] writable_roots = [ "/tmp" ] [sandbox_danger_full_access] # danger-full-access options here... ``` In this scheme, you never need to comment out the configuration for an individual sandbox type: you only need to redefine `sandbox_mode`. Relatedly, previous to this change, a user had to do `-c sandbox.mode=read-only` to change the mode on the command line. With this change, things are arguably a bit cleaner because the equivalent option is `-c sandbox_mode=read-only` (and now `-c sandbox_workspace_write=...` can be set separately). Though more importantly, we introduce the `-s/--sandbox` option to the CLI, which maps directly to `sandbox_mode` in `config.toml`, making config override behavior easier to reason about. Moreover, as you can see in the updates to the various Markdown files, it is much easier to explain how to configure sandboxing when things like `--sandbox read-only` can be used as an example. Relatedly, this cleanup also made it straightforward to add support for a `sandbox` option for Codex when used as an MCP server (see the changes to `mcp-server/src/codex_tool_config.rs`). Fixes https://github.com/openai/codex/issues/1248.
2025-07-07 22:31:30 -07:00
"sandbox": {
"description": "Sandbox mode: `read-only`, `workspace-write`, or `danger-full-access`.",
"enum": [
"read-only",
"workspace-write",
"danger-full-access"
],
"type": "string"
},
feat: add support for -c/--config to override individual config items (#1137) This PR introduces support for `-c`/`--config` so users can override individual config values on the command line using `--config name=value`. Example: ``` codex --config model=o4-mini ``` Making it possible to set arbitrary config values on the command line results in a more flexible configuration scheme and makes it easier to provide single-line examples that can be copy-pasted from documentation. Effectively, it means there are four levels of configuration for some values: - Default value (e.g., `model` currently defaults to `o4-mini`) - Value in `config.toml` (e.g., user could override the default to be `model = "o3"` in their `config.toml`) - Specifying `-c` or `--config` to override `model` (e.g., user can include `-c model=o3` in their list of args to Codex) - If available, a config-specific flag can be used, which takes precedence over `-c` (e.g., user can specify `--model o3` in their list of args to Codex) Now that it is possible to specify anything that could be configured in `config.toml` on the command line using `-c`, we do not need to have a custom flag for every possible config option (which can clutter the output of `--help`). To that end, as part of this PR, we drop support for the `--disable-response-storage` flag, as users can now specify `-c disable_response_storage=true` to get the equivalent functionality. Under the hood, this works by loading the `config.toml` into a `toml::Value`. Then for each `key=value`, we create a small synthetic TOML file with `value` so that we can run the TOML parser to get the equivalent `toml::Value`. We then parse `key` to determine the point in the original `toml::Value` to do the insert/replace. Once all of the overrides from `-c` args have been applied, the `toml::Value` is deserialized into a `ConfigToml` and then the `ConfigOverrides` are applied, as before.
2025-05-27 23:11:44 -07:00
"config": {
"description": "Individual config settings that will override what is in CODEX_HOME/config.toml.",
"additionalProperties": true,
"type": "object"
},
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
"cwd": {
"description": "Working directory for the session. If relative, it is resolved against the server process's current working directory.",
"type": "string"
},
"include-plan-tool": {
"description": "Whether to include the plan tool in the conversation.",
"type": "boolean"
},
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
"model": {
"description": "Optional override for the model name (e.g. \"o3\", \"o4-mini\").",
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
"type": "string"
},
"profile": {
"description": "Configuration profile from config.toml to specify default options.",
"type": "string"
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
},
"prompt": {
"description": "The *initial user prompt* to start the Codex conversation.",
"type": "string"
},
"base-instructions": {
"description": "The set of instructions to use instead of the default ones.",
"type": "string"
},
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
},
"required": [
"prompt"
]
}
});
assert_eq!(expected_tool_json, tool_json);
}
#[test]
fn verify_codex_tool_reply_json_schema() {
let tool = create_tool_for_codex_tool_call_reply_param();
let tool_json = serde_json::to_value(&tool).expect("tool serializes");
let expected_tool_json = serde_json::json!({
"description": "Continue a Codex conversation by providing the conversation id and prompt.",
"inputSchema": {
"properties": {
"conversationId": {
"description": "The conversation id for this Codex session.",
"type": "string"
},
"prompt": {
"description": "The *next user prompt* to continue the Codex conversation.",
"type": "string"
},
},
"required": [
"conversationId",
"prompt",
],
"type": "object",
},
"name": "codex-reply",
"title": "Codex Reply",
});
assert_eq!(expected_tool_json, tool_json);
}
feat: make Codex available as a tool when running it as an MCP server (#811) This PR replaces the placeholder `"echo"` tool call in the MCP server with a `"codex"` tool that calls Codex. Events such as `ExecApprovalRequest` and `ApplyPatchApprovalRequest` are not handled properly yet, but I have `approval_policy = "never"` set in my `~/.codex/config.toml` such that those codepaths are not exercised. The schema for this MPC tool is defined by a new `CodexToolCallParam` struct introduced in this PR. It is fairly similar to `ConfigOverrides`, as the param is used to help create the `Config` used to start the Codex session, though it also includes the `prompt` used to kick off the session. This PR also introduces the use of the third-party `schemars` crate to generate the JSON schema, which is verified in the `verify_codex_tool_json_schema()` unit test. Events that are dispatched during the Codex session are sent back to the MCP client as MCP notifications. This gives the client a way to monitor progress as the tool call itself may take minutes to complete depending on the complexity of the task requested by the user. In the video below, I launched the server via: ```shell mcp-server$ RUST_LOG=debug npx @modelcontextprotocol/inspector cargo run -- ``` In the video, you can see the flow of: * requesting the list of tools * choosing the **codex** tool * entering a value for **prompt** and then making the tool call Note that I left the other fields blank because when unspecified, the values in my `~/.codex/config.toml` were used: https://github.com/user-attachments/assets/1975058c-b004-43ef-8c8d-800a953b8192 Note that while using the inspector, I did run into https://github.com/modelcontextprotocol/inspector/issues/293, though the tip about ensuring I had only one instance of the **MCP Inspector** tab open in my browser seemed to fix things.
2025-05-05 07:16:19 -07:00
}