Files
codex/codex-rs/tools
sayan-oai 0df7e9a820 register all mcp tools with namespace (#17404)
stacked on #17402.

MCP tools returned by `tool_search` (deferred tools) get registered in
our `ToolRegistry` with a different format than directly available
tools. this leads to two different ways of accessing MCP tools from our
tool catalog, only one of which works for each. fix this by registering
all MCP tools with the namespace format, since this info is already
available.

also, direct MCP tools are registered to responsesapi without a
namespace, while deferred MCP tools have a namespace. this means we can
receive MCP `FunctionCall`s in both formats from namespaces. fix this by
always registering MCP tools with namespace, regardless of deferral
status.

make code mode track `ToolName` provenance of tools so it can map the
literal JS function name string to the correct `ToolName` for
invocation, rather than supporting both in core.

this lets us unify to a single canonical `ToolName` representation for
each MCP tool and force everywhere to use that one, without supporting
fallbacks.
2026-04-15 21:02:59 +08:00
..

codex-tools

codex-tools is intended to become the home for tool-related code that is shared across multiple crates and does not need to stay coupled to codex-core.

Today this crate is intentionally small. It currently owns the shared tool schema and Responses API tool primitives that no longer need to live in core/src/tools/spec.rs or core/src/client_common.rs:

  • JsonSchema
  • AdditionalProperties
  • ToolDefinition
  • ToolSpec
  • ConfiguredToolSpec
  • ResponsesApiTool
  • FreeformTool
  • FreeformToolFormat
  • ToolSearchOutputTool
  • ResponsesApiWebSearchFilters
  • ResponsesApiWebSearchUserLocation
  • ResponsesApiNamespace
  • ResponsesApiNamespaceTool
  • code-mode ToolSpec adapters and exec / wait spec builders
  • JS REPL spec builders
  • MCP resource, list_dir, and test_sync_tool spec builders
  • local host tool spec builders for shell/exec/request-permissions/view-image
  • collaboration and agent-job ToolSpec builders for spawn/send/wait/close, request_user_input, and CSV fanout/reporting
  • discoverable-tool models, client filtering, and ToolSpec builders for tool_search and tool_suggest
  • parse_tool_input_schema()
  • parse_dynamic_tool()
  • parse_mcp_tool()
  • create_tools_json_for_responses_api()
  • mcp_call_tool_result_output_schema()
  • tool_definition_to_responses_api_tool()
  • dynamic_tool_to_responses_api_tool()
  • mcp_tool_to_responses_api_tool()
  • mcp_tool_to_deferred_responses_api_tool()
  • augment_tool_spec_for_code_mode()
  • tool_spec_to_code_mode_tool_definition()

That extraction is the first step in a longer migration. The goal is not to move all of core/src/tools into this crate in one shot. Instead, the plan is to peel off reusable pieces in reviewable increments while keeping compatibility-sensitive orchestration in codex-core until the surrounding boundaries are ready.

Vision

Over time, this crate should hold tool-facing primitives that are shared by multiple consumers, for example:

  • schema and spec data models
  • tool input/output parsing helpers
  • tool metadata and compatibility shims that do not depend on codex-core
  • other narrowly scoped utility code that multiple crates need

The corresponding non-goals are just as important:

  • do not move codex-core orchestration here prematurely
  • do not pull Session / TurnContext / approval flow / runtime execution logic into this crate unless those dependencies have first been split into stable shared interfaces
  • do not turn this crate into a grab-bag for unrelated helper code

Migration approach

The expected migration shape is:

  1. Move low-coupling tool primitives here.
  2. Switch non-core consumers to depend on codex-tools directly.
  3. Leave compatibility-sensitive adapters in codex-core while downstream call sites are updated.
  4. Only extract higher-level tool infrastructure after the crate boundaries are clear and independently testable.

That means it is normal for codex-core to temporarily re-export types or helpers from codex-tools during the transition.

Crate conventions

This crate should start with stricter structure than core/src/tools so it stays easy to grow:

  • src/lib.rs should remain exports-only.
  • Business logic should live in named module files such as foo.rs.
  • Unit tests for foo.rs should live in a sibling foo_tests.rs.
  • The implementation file should wire tests with:
#[cfg(test)]
#[path = "foo_tests.rs"]
mod tests;

If this crate starts accumulating code that needs runtime state from codex-core, that is a sign to revisit the extraction boundary before adding more here.