feat(mcp): supports_parallel_tool_calls for MCP servers (salvage of #9944)#26825
Merged
Conversation
Port from openai/codex#17667: MCP servers can now opt-in to parallel tool execution by setting supports_parallel_tool_calls: true in their config. This allows tools from the same server to run concurrently within a single tool-call batch, matching the behavior already available for built-in tools like web_search and read_file. Previously all MCP tools were forced sequential because they weren't in the _PARALLEL_SAFE_TOOLS set. Now _should_parallelize_tool_batch checks is_mcp_tool_parallel_safe() which looks up the server's config flag. Config example: mcp_servers: docs: command: "docs-server" supports_parallel_tool_calls: true Changes: - tools/mcp_tool.py: Track parallel-safe servers in _parallel_safe_servers set, populated during register_mcp_servers(). Add is_mcp_tool_parallel_safe() public API. - run_agent.py: Add _is_mcp_tool_parallel_safe() lazy-import wrapper. Update _should_parallelize_tool_batch() to check MCP tools against server config. - 11 new tests covering the feature end-to-end. - Updated MCP docs and config reference.
Contributor
🔎 Lint report:
|
1 task
DIZ-admin
pushed a commit
to DIZ-admin/hermes-agent
that referenced
this pull request
May 16, 2026
…26825) Port from openai/codex#17667: MCP servers can now opt-in to parallel tool execution by setting supports_parallel_tool_calls: true in their config. This allows tools from the same server to run concurrently within a single tool-call batch, matching the behavior already available for built-in tools like web_search and read_file. Previously all MCP tools were forced sequential because they weren't in the _PARALLEL_SAFE_TOOLS set. Now _should_parallelize_tool_batch checks is_mcp_tool_parallel_safe() which looks up the server's config flag. Config example: mcp_servers: docs: command: "docs-server" supports_parallel_tool_calls: true Changes: - tools/mcp_tool.py: Track parallel-safe servers in _parallel_safe_servers set, populated during register_mcp_servers(). Add is_mcp_tool_parallel_safe() public API. - run_agent.py: Add _is_mcp_tool_parallel_safe() lazy-import wrapper. Update _should_parallelize_tool_batch() to check MCP tools against server config. - 11 new tests covering the feature end-to-end. - Updated MCP docs and config reference.
venyon2k
pushed a commit
to venyon2k/hermes-agent
that referenced
this pull request
May 17, 2026
…26825) Port from openai/codex#17667: MCP servers can now opt-in to parallel tool execution by setting supports_parallel_tool_calls: true in their config. This allows tools from the same server to run concurrently within a single tool-call batch, matching the behavior already available for built-in tools like web_search and read_file. Previously all MCP tools were forced sequential because they weren't in the _PARALLEL_SAFE_TOOLS set. Now _should_parallelize_tool_batch checks is_mcp_tool_parallel_safe() which looks up the server's config flag. Config example: mcp_servers: docs: command: "docs-server" supports_parallel_tool_calls: true Changes: - tools/mcp_tool.py: Track parallel-safe servers in _parallel_safe_servers set, populated during register_mcp_servers(). Add is_mcp_tool_parallel_safe() public API. - run_agent.py: Add _is_mcp_tool_parallel_safe() lazy-import wrapper. Update _should_parallelize_tool_batch() to check MCP tools against server config. - 11 new tests covering the feature end-to-end. - Updated MCP docs and config reference.
teknium1
added a commit
that referenced
this pull request
May 17, 2026
…t to tool_dispatch_helpers Original commit 395e9dd by Teknium targeted module-level _is_mcp_tool_parallel_safe and _should_parallelize_tool_batch helpers in pre-refactor run_agent.py. Both helpers now live in agent/tool_dispatch_helpers.py — re-applied to that module. The tools/mcp_tool.py portion (the public is_mcp_tool_parallel_safe API + _parallel_safe_servers tracking) merged cleanly from main via the prior merge commit. Co-authored-by: Teknium <127238744+teknium1@users.noreply.github.com>
gweeteve
pushed a commit
to gweeteve/hermes-agent
that referenced
this pull request
Jun 2, 2026
…26825) Port from openai/codex#17667: MCP servers can now opt-in to parallel tool execution by setting supports_parallel_tool_calls: true in their config. This allows tools from the same server to run concurrently within a single tool-call batch, matching the behavior already available for built-in tools like web_search and read_file. Previously all MCP tools were forced sequential because they weren't in the _PARALLEL_SAFE_TOOLS set. Now _should_parallelize_tool_batch checks is_mcp_tool_parallel_safe() which looks up the server's config flag. Config example: mcp_servers: docs: command: "docs-server" supports_parallel_tool_calls: true Changes: - tools/mcp_tool.py: Track parallel-safe servers in _parallel_safe_servers set, populated during register_mcp_servers(). Add is_mcp_tool_parallel_safe() public API. - run_agent.py: Add _is_mcp_tool_parallel_safe() lazy-import wrapper. Update _should_parallelize_tool_batch() to check MCP tools against server config. - 11 new tests covering the feature end-to-end. - Updated MCP docs and config reference.
gweeteve
pushed a commit
to gweeteve/hermes-agent
that referenced
this pull request
Jun 2, 2026
…26825) — port to tool_dispatch_helpers Original commit 395e9dd by Teknium targeted module-level _is_mcp_tool_parallel_safe and _should_parallelize_tool_batch helpers in pre-refactor run_agent.py. Both helpers now live in agent/tool_dispatch_helpers.py — re-applied to that module. The tools/mcp_tool.py portion (the public is_mcp_tool_parallel_safe API + _parallel_safe_servers tracking) merged cleanly from main via the prior merge commit. Co-authored-by: Teknium <127238744+teknium1@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Salvage of #9944 — MCP servers can opt-in to parallel tool execution by setting
supports_parallel_tool_calls: truein their config. Tools from opted-in servers can run concurrently within a single tool-call batch, matching the behavior already available for read-only built-ins likeweb_searchandread_file.Port from openai/codex#17667.
Why salvage
Original PR is 4,266 commits stale. Cherry-picking onto current main hit one conflict in
tools/mcp_tool.py(circuit breaker module-level state block was added on main between the PR's branch point and today) — resolved by keeping main's block and appending the PR's_parallel_safe_serversset. Also fixed two tests that mocked_sync_mcp_toolsets, a helper that no longer exists on main.Changes
tools/mcp_tool.py:_parallel_safe_serversmodule-level set, populated duringregister_mcp_servers()(idempotent, handles toggling). New publicis_mcp_tool_parallel_safe(tool_name)walks registered prefixes (server names can contain underscores after sanitization).run_agent.py: lazy-import wrapper_is_mcp_tool_parallel_safe()._should_parallelize_tool_batch()now consults it for MCP tools that aren't in the static_PARALLEL_SAFE_TOOLSset._should_parallelize_tool_batch.mcp.md+mcp-config-reference.mdupdated.Validation
tests/tools/test_mcp_tool.pytests/run_agent/test_run_agent.py+ parallel MCP testsArchitectural note
Codex implements this through
ToolRouterat the Rust crate level. We use a module-level set populated during server registration and queried viais_mcp_tool_parallel_safe(tool_name), which handles underscores-in-server-names by checking all registered parallel-safe prefixes. Same external behavior, different glue.Credit
@teknium1's original scout work in #9944. This PR salvages it onto current main.