fix(code): run MCP login during a run, queue the restart - #4643
Merged
Mason Daugherty (mdrxy) merged 6 commits intoJul 14, 2026
Conversation
Previously `/mcp login` deferred the entire OAuth flow until the active agent/shell task finished. Only the final server restart actually conflicts with a live run, so start login immediately and queue just the restart via a new `mcp_reconnect` deferred action that drains once idle. Co-authored-by: open-swe[bot] <open-swe@users.noreply.github.com>
Mason Daugherty (mdrxy)
marked this pull request as ready for review
July 10, 2026 21:40
Mason Daugherty (mdrxy)
marked this pull request as draft
July 13, 2026 13:35
Mason Daugherty (mdrxy)
marked this pull request as ready for review
July 14, 2026 03:30
Mason Daugherty (mdrxy)
deleted the
mdrxy/code/mcp-login-during-generation
branch
July 14, 2026 04:12
Mason Daugherty (mdrxy)
added a commit
that referenced
this pull request
Jul 14, 2026
> [!CAUTION] > Merging this PR will automatically publish to **PyPI** and create a **GitHub release**. For the full release process, see [`.github/RELEASING.md`](https://github.com/langchain-ai/deepagents/blob/main/.github/RELEASING.md). --- _Release notes preview: keep this section in sync with the package `CHANGELOG.md`. The published GitHub release body is extracted from the merged `CHANGELOG.md` by `release.yml`, not from this PR description._ --- ## [0.1.38](deepagents-code==0.1.37...deepagents-code==0.1.38) (2026-07-14) ### Features * Improve `/goal` criteria UX ([#4694](#4694)) ([06f46ff](06f46ff)) * Add interactive goal management ([#4693](#4693)) ([64205e2](64205e2)) * Conditionally drop `TodoListMiddleware` ([#4685](#4685)) ([d3a3077](d3a3077)) * Add `memory.auto_save` config flag ([#4700](#4700)) ([55b60ca](55b60ca)) * Trace experimental mode in metadata ([#4705](#4705)) ([22d5045](22d5045)) ### Bug Fixes * Align context diff rows with changed rows ([#4714](#4714)) ([f9915db](f9915db)) * Hydrate virtualized transcript on scroll offset changes ([#4646](#4646)) ([f77eeb0](f77eeb0)) * Keep TODO and edit tools expanded ([#4704](#4704)) ([1d549d3](1d549d3)) * Make `/goal` completion and grading reliable ([#4691](#4691)) ([9da63c6](9da63c6)) * Preflight Ollama host reachability before discovery probe ([#4702](#4702)) ([96fe71a](96fe71a)) * Quiet expected non-repo `git ls-files` logging ([#4701](#4701)) ([3d499db](3d499db)) * Rename `ls_agent_kind` to `ls_agent_purpose` ([#4708](#4708)) ([2678b16](2678b16)) * Run MCP login during a run, queue the restart ([#4643](#4643)) ([65e1ee8](65e1ee8)) * Show "Took <duration>" when `task` tool completes ([#4638](#4638)) ([19538e0](19538e0)) * Support unambiguous `read_file` gutters ([#4711](#4711)) ([2089b54](2089b54)) --- _End release notes preview._ --- > [!NOTE] > A **New Contributors** section is appended to the GitHub release notes automatically at publish time (see [Release Pipeline](https://github.com/langchain-ai/deepagents/blob/main/.github/RELEASING.md#release-pipeline), step 2). --------- Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: Mason Daugherty <github@mdrxy.com>
Marcelo5444
pushed a commit
to Marcelo5444/deepagents
that referenced
this pull request
Jul 30, 2026
…i#4643) `/mcp login` now runs during an active generation instead of waiting for it to finish; if you accept the reconnect, the server restart is queued and applied automatically once the current task completes. --- `/mcp login` previously deferred the *entire* OAuth flow until an active agent or shell task finished. But only the final server restart (`_restart_server_for_mcp_refresh` → `_respawn_server`) actually conflicts with a live run — the OAuth handshake and on-disk token write never touch the running server. This starts login immediately even during a generation and queues only the restart, so the user can authenticate without waiting for the turn to end. When the user accepts "reconnect" on the post-login prompt while a run is active, the restart is queued as a new `mcp_reconnect` deferred action (drained by the existing `_maybe_drain_deferred` on agent/shell cleanup) instead of tearing down the in-flight generation. The reconnect is marked pending in the meantime so `/mcp reconnect` and the splash banner stay accurate, and the queued action re-checks the pending flag to avoid a redundant restart if a manual `/mcp reconnect` already ran. When idle, behavior is unchanged (immediate restart). Made by [Open SWE](https://openswe.vercel.app/agents/4defa6fb-fe26-d47b-9fa0-10b5a9c29dcd) --------- Co-authored-by: open-swe[bot] <open-swe@users.noreply.github.com>
Marcelo5444
pushed a commit
to Marcelo5444/deepagents
that referenced
this pull request
Jul 30, 2026
> [!CAUTION] > Merging this PR will automatically publish to **PyPI** and create a **GitHub release**. For the full release process, see [`.github/RELEASING.md`](https://github.com/langchain-ai/deepagents/blob/main/.github/RELEASING.md). --- _Release notes preview: keep this section in sync with the package `CHANGELOG.md`. The published GitHub release body is extracted from the merged `CHANGELOG.md` by `release.yml`, not from this PR description._ --- ## [0.1.38](langchain-ai/deepagents@deepagents-code==0.1.37...deepagents-code==0.1.38) (2026-07-14) ### Features * Improve `/goal` criteria UX ([langchain-ai#4694](langchain-ai#4694)) ([06f46ff](langchain-ai@06f46ff)) * Add interactive goal management ([langchain-ai#4693](langchain-ai#4693)) ([64205e2](langchain-ai@64205e2)) * Conditionally drop `TodoListMiddleware` ([langchain-ai#4685](langchain-ai#4685)) ([d3a3077](langchain-ai@d3a3077)) * Add `memory.auto_save` config flag ([langchain-ai#4700](langchain-ai#4700)) ([55b60ca](langchain-ai@55b60ca)) * Trace experimental mode in metadata ([langchain-ai#4705](langchain-ai#4705)) ([22d5045](langchain-ai@22d5045)) ### Bug Fixes * Align context diff rows with changed rows ([langchain-ai#4714](langchain-ai#4714)) ([f9915db](langchain-ai@f9915db)) * Hydrate virtualized transcript on scroll offset changes ([langchain-ai#4646](langchain-ai#4646)) ([f77eeb0](langchain-ai@f77eeb0)) * Keep TODO and edit tools expanded ([langchain-ai#4704](langchain-ai#4704)) ([1d549d3](langchain-ai@1d549d3)) * Make `/goal` completion and grading reliable ([langchain-ai#4691](langchain-ai#4691)) ([9da63c6](langchain-ai@9da63c6)) * Preflight Ollama host reachability before discovery probe ([langchain-ai#4702](langchain-ai#4702)) ([96fe71a](langchain-ai@96fe71a)) * Quiet expected non-repo `git ls-files` logging ([langchain-ai#4701](langchain-ai#4701)) ([3d499db](langchain-ai@3d499db)) * Rename `ls_agent_kind` to `ls_agent_purpose` ([langchain-ai#4708](langchain-ai#4708)) ([2678b16](langchain-ai@2678b16)) * Run MCP login during a run, queue the restart ([langchain-ai#4643](langchain-ai#4643)) ([65e1ee8](langchain-ai@65e1ee8)) * Show "Took <duration>" when `task` tool completes ([langchain-ai#4638](langchain-ai#4638)) ([19538e0](langchain-ai@19538e0)) * Support unambiguous `read_file` gutters ([langchain-ai#4711](langchain-ai#4711)) ([2089b54](langchain-ai@2089b54)) --- _End release notes preview._ --- > [!NOTE] > A **New Contributors** section is appended to the GitHub release notes automatically at publish time (see [Release Pipeline](https://github.com/langchain-ai/deepagents/blob/main/.github/RELEASING.md#release-pipeline), step 2). --------- Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: Mason Daugherty <github@mdrxy.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
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.
/mcp loginnow runs during an active generation instead of waiting for it to finish; if you accept the reconnect, the server restart is queued and applied automatically once the current task completes./mcp loginpreviously deferred the entire OAuth flow until an active agent or shell task finished. But only the final server restart (_restart_server_for_mcp_refresh→_respawn_server) actually conflicts with a live run — the OAuth handshake and on-disk token write never touch the running server. This starts login immediately even during a generation and queues only the restart, so the user can authenticate without waiting for the turn to end.When the user accepts "reconnect" on the post-login prompt while a run is active, the restart is queued as a new
mcp_reconnectdeferred action (drained by the existing_maybe_drain_deferredon agent/shell cleanup) instead of tearing down the in-flight generation. The reconnect is marked pending in the meantime so/mcp reconnectand the splash banner stay accurate, and the queued action re-checks the pending flag to avoid a redundant restart if a manual/mcp reconnectalready ran. When idle, behavior is unchanged (immediate restart).Made by Open SWE