[Agent Builder] Add todo list tool and per-round TodosStep UI by KDKHD · Pull Request #265578 · elastic/kibana

KDKHD · 2026-04-24T15:49:45Z

Summary

Part of https://github.com/elastic/search-team/issues/13832

Screen.Recording.2026-04-24.at.17.39.10.mov

Adds a new platform_core_todo_write built-in tool that lets the research agent plan and track work during a conversation round
Introduces TodosStep as a member of ConversationRoundStep so each round permanently records its todo snapshot — visible when scrolling through conversation history
Incomplete todos carry over automatically from the previous round; if all todos completed, nothing carries over
Frontend collapses carried-over todos to a "To-dos N" header and expands with a CSS animation when the agent calls TodoWrite during the round
Tool and system prompt are gated behind the existing agentBuilder:experimentalFeatures UI setting

Changes

Common packages

agent-builder-common: Added TodosStep, isTodosStep, TodosStepData (with carried_over flag) to the ConversationRoundStep union; exported todo tool constants
agent-builder-server: Added todos: boolean to ExperimentalFeatures; new TodoStateManager (in-memory per-execution, initialized from conversation state); exported from runner index

Server / execution

New built-in tool server/services/tools/builtin/todo/todo_write.ts — replaces the full todo list on each call
select_tools.ts: todo tool gated behind experimentalFeatures.todos
run_chat_agent.ts: captures initialTodos before the round starts, passes to addRoundCompleteEvent
add_round_complete_event.ts: uses initialTodos as fallback TodosStep with carried_over: true when agent never called TodoWrite; helper carriedOverTodos skips carryover when all todos are done
run_agent.ts / runner.ts: wires todos experimental flag and todoStateManager
research_agent.ts: includes getTodoInstructions() prompt section when experimentalFeatures.todos is enabled
New prompts/utils/todos.ts: opencode-inspired task management prompt explaining when/how to use the tool, including the full-list replacement semantic

Frontend

New todos_step_display.tsx: collapsed header-only view when carried_over, full expanded list with expandIn keyframe animation on mount when not carried over
round_layout.tsx: renders TodosStepDisplay per-round between the thinking section and the response
use_conversation_actions.ts: addOptimisticRound seeds the new round with carried-over todos; new addOrUpdateTodosStep action flips carried_over → false triggering the expand animation
use_subscribe_to_chat_events.ts: wired to call addOrUpdateTodosStep on todo events
types.ts (conversation client): TodosStep added to PersistentConversationRoundStep for ES serialization

Tests

chat_message_text.test.tsx: added addOrUpdateTodosStep: jest.fn() to conversation actions mock
test_utils/runner.ts: added todos: false to experimentalFeatures mock

Test plan

Enable agentBuilder:experimentalFeatures in Kibana advanced settings
Start a new conversation with a multi-step task; confirm agent creates todos and they appear in the round. e.g. "Can you please search for security alerts with these keywords (using security.alerts), one by one: Virus, malware, macOS. Create a to-do list for each."
Start a second round with incomplete todos (by sending something like "Add windows to the list"); confirm they carry over collapsed ("To-dos N"), then expand when agent calls TodoWrite
Start a new round that does not modify todos (e.g. say "Thanks") - confirm todos were carried over
Ask the agent to "Complete the todos sequentially" - observe all todos being marked as completed
Say "Thanks" to the agent - observer todos are not carried over because they are all complete
Refresh the page; confirm the todo state and collapsed/expanded display is preserved from persisted round data
Disable the experimental setting; confirm the todo tool is not available, and the todo prompt section is not part of the prompt anymore (check inside of Langsmith or other observability tooling)

Implements a persistent task management feature for the AI agent: - New `platform_core_todo_write` built-in tool that lets the research agent create and update a todo list during a conversation round - `TodoStateManager` tracks in-memory todo state per execution, initialized from the conversation's persisted state - `TodosStep` added to `ConversationRoundStep` union so each round carries its own snapshot of todos, visible when scrolling through history - Incomplete todos are carried over from the previous round at the start of each new round (`carried_over: true`); if all todos are complete, nothing is carried over - Frontend: `TodosStepDisplay` renders carried-over todos as a collapsed "To-dos N" header; once the agent calls `TodoWrite` the component expands with a CSS animation - `addOrUpdateTodosStep` in `use_conversation_actions` handles both the optimistic carryover on round start and the live expansion when the tool fires - Tool and system prompt (based on opencode's task management instructions) are gated behind the `agentBuilder:experimentalFeatures` UI setting - `todos: false` added to `experimentalFeatures` in test mocks

- Adds explicit rule that marking a task in_progress and completing the previous can happen in the same TodoWrite call - Carries over todo_panel.tsx and use_todo_list.ts (unused legacy files, kept for reference)

Conflict resolution: - x-pack/platform/plugins/shared/agent_builder/server/services/execution/run_agent/utils/select_tools.ts: Combined the upstream `Promise<SelectToolsResult>` return type and the new per-tool `withOrigin(...)` tagging with the branch's `todoStateManager` parameter and `todoTools` block. Todo tools are tagged `ToolOrigin.internal` to match the convention used for the similar builtin filestore tools. Follow-up adjustment from the resolution: - x-pack/platform/plugins/shared/agent_builder/server/services/execution/run_agent/utils/select_tools.test.ts (new file from upstream): Pass a `todoStateManager` mock to the two `selectTools` invocations so the newly required parameter type-checks. Verified locally: type_check (agent_builder plugin, agent-builder-server, agent-builder-common) green; jest (select_tools, chat_message_text) green; eslint clean on touched files.

chrisbmar

(frontend files review only)

Nice work Kenneth @KDKHD 🚀 super neat feature and something that will bring big benefit, especially after we have design input. Looking forward to testing this out further.

How do you see this growing in the future? Where do you envision this going to (just off the top of your head what does a good todo list experience look like to you and what are the low hanging fruits)? I think the frontend part of it is fine at the moment with the carryover logic into the next optimistic round, but I am conscious of this growing in complexity on the FE.

FYI one of us will have a conflict depending on who's PR gets merged first yours, or mine. I can help you resolve the conflict if mine gets merged first. It refactors all the streaming logic. But anyway, it should be 1 or 2 file changes and just a few lines, Claude could handle it with ease.

chrisbmar · 2026-05-06T09:59:47Z

+/**
+ * Returns the todo list to carry over from the previous round, or undefined if nothing should carry over.
+ * Carryover only happens when at least one item is still incomplete (pending / in_progress).
+ * When carried over, both complete and incomplete items are included so the full plan is visible.
+ */
+const carriedOverTodos = (todos: TodoItem[] | undefined): TodoItem[] | undefined => {
+  if (!todos?.length) return undefined;
+  const hasIncomplete = todos.some((t) => t.status !== 'completed' && t.status !== 'cancelled');
+  return hasIncomplete ? todos : undefined;
+};


It would be nice to move this to somewhere shared and use it on the frontend too in use_conversation_actions so that it's not duplicated and so that there is no future drift.

pgayvallet

Looking good. A few comments and nits (i'll be faster for the second review I promise)

pgayvallet · 2026-05-06T07:40:00Z

@@ -78,6 +77,7 @@ export enum ConversationRoundStepType {
  reasoning = 'reasoning',
  compaction = 'compaction',
  backgroundAgentComplete = 'background_agent_complete',
+  todos = 'todos',


NIT: I would rename that to update_todos to be slightly more explicit

pgayvallet · 2026-05-06T07:48:58Z

+export const todoTools = {
+  write: platformCoreTool('todo_write'),
+} as const;


Let's add it to the internalTools map instead, we have too many different records already (and we can just call it write_todos - no need for prefixing)

pgayvallet · 2026-05-06T09:11:24Z

+export type TodoStatus = 'pending' | 'in_progress' | 'completed' | 'cancelled';
+export type TodoPriority = 'high' | 'medium' | 'low';
+
+export interface TodoItem {
+  content: string;
+  status: TodoStatus;
+  priority: TodoPriority;
+}


Given tasks are naturally ordered, I wonder if the priority really matters, especially given it's not surfaced in the UI. Do you think it's still worth having, or should we just remove it for now?

Removed it!

pgayvallet · 2026-05-06T09:15:53Z

+import { useState } from 'react';
+import type { TodoItem } from '@kbn/agent-builder-common/chat/conversation';
+
+export const useTodoList = (initialTodos?: TodoItem[]) => {
+  const [todos, setTodos] = useState<TodoItem[]>(initialTodos ?? []);
+  return { todos, setTodos };
+};


This looks unused

pgayvallet · 2026-05-06T09:22:10Z

+
+export const getTodoInstructions = (): string => `## TASK MANAGEMENT
+
+You have access to the TodoWrite tool to plan and track your work. Use it frequently so the user can see what you are doing and how you are progressing.


need to replace TodoWrite with the actual tool name here (other references in the prompt too)

pgayvallet · 2026-05-06T09:37:30Z

+export const TodoPanel: React.FC<TodoPanelProps> = ({ todos }) => {
+  if (todos.length === 0) return null;
+
+  const grouped = PRIORITY_ORDER.reduce<Record<TodoPriority, TodoItem[]>>(
+    (acc, p) => {
+      acc[p] = todos.filter((t) => t.priority === p);
+      return acc;
+    },
+    { high: [], medium: [], low: [] }
+  );


This (and so the whole file) looks unused

pgayvallet · 2026-05-06T09:39:50Z

+          const prevTodosStep = draft?.rounds?.at(-1)?.steps?.filter(isTodosStep)?.at(-1);
+          const carryoverTodos = prevTodosStep?.todos?.some(
+            (t) => t.status !== 'completed' && t.status !== 'cancelled'
+          )
+            ? prevTodosStep.todos
+            : undefined;


NIT: extract as util /move to same place isTodosStep lives?

pgayvallet · 2026-05-06T09:56:14Z

+  tags: ['internal'],
+  handler: async ({ todos }, context) => {
+    todoStateManager.set(todos);
+    context.events.sendUiEvent('todos_updated', { todos });


(just thinking) so I don't think we really needed this, because we could have have listened to tool calls events for that particular tool directly instead (in use_subscribe_to_chat_events.ts), but this works too.

pgayvallet · 2026-05-06T10:04:27Z

@@ -286,6 +295,14 @@ const createRound = ({
  const thinkingCompleteEvent = events.find(isThinkingCompleteEvent);
  const promptRequestEvents = events.filter(isPromptRequestEvent);

+  // Collect todos_updated UI events; only the last snapshot is stored as a round step
+  const lastTodosData = events.reduce<TodoItem[] | undefined>((last, e) => {
+    if (isToolUiEvent<'todos_updated', { todos: TodoItem[] }>(e, 'todos_updated')) {


the todos_updated ui event name is used in various places, let's have a const for it

pgayvallet · 2026-05-06T10:05:34Z

+  const carriedTodos = carriedOverTodos(initialTodos);
+  const todosForStep = lastTodosData ?? carriedTodos;


NIT: avoid unnecessary call

const todosForStep = lastTodosData ?? carriedOverTodos(initialTodos);

…bana into feature-agent-builder-todo-list

…er-todo-list # Conflicts: # x-pack/platform/packages/shared/agent-builder/agent-builder-common/index.ts # x-pack/platform/plugins/shared/agent_builder/server/services/execution/run_agent/prompts/answer_agent.ts

…bana into feature-agent-builder-todo-list

KDKHD · 2026-05-07T11:40:44Z

/ci

pgayvallet

LGTM

KDKHD · 2026-05-11T10:08:12Z

@elasticmachine merge upstream

chrisbmar

lets gooo

kibanamachine · 2026-05-11T11:17:09Z

💛 Build succeeded, but was flaky

Buildkite Build
Commit: 7b2cc55

Failed CI Steps

Test Failures

[job] [logs] Scout Lane #8 - stateful-classic / default / local-stateful-classic - APM integration not installed but setup completed - Admin user
[job] [logs] Scout Lane #8 - stateful-classic / default / local-stateful-classic - Profiling is not setup and no data is loaded - Admin users
[job] [logs] Scout Lane #8 - stateful-classic / default / local-stateful-classic - Profiling is not setup and no data is loaded - Viewer users

Metrics [docs]

Module Count

Fewer modules leads to a faster build time

id	before	after	diff
`agentBuilder`	1528	1529	+1

Async chunks

Total size of all lazy-loaded chunks that will be downloaded as the user navigates the app

id	before	after	diff
`agentBuilder`	1.1MB	1.1MB	+3.2KB

Unknown metric groups

API count

id	before	after	diff
`@kbn/agent-builder-common`	727	740	+13
`@kbn/agent-builder-server`	529	531	+2
total			+15

History

💚 Build #439368 succeeded 70d9415
💛 Build #438842 was flaky ab059be
💔 Build #436457 failed 08cda1e

## Summary Part of elastic/search-team#13832 <img width="960" height="702" alt="image" src="https://github.com/user-attachments/assets/2995f47e-7f33-463f-a426-2c07c337c811" /> https://github.com/user-attachments/assets/13cd178c-2f8c-4d30-afea-75ef460a2125 - Adds a new `platform_core_todo_write` built-in tool that lets the research agent plan and track work during a conversation round - Introduces `TodosStep` as a member of `ConversationRoundStep` so each round permanently records its todo snapshot — visible when scrolling through conversation history - Incomplete todos carry over automatically from the previous round; if all todos completed, nothing carries over - Frontend collapses carried-over todos to a "To-dos N" header and expands with a CSS animation when the agent calls `TodoWrite` during the round - Tool and system prompt are gated behind the existing `agentBuilder:experimentalFeatures` UI setting ## Changes **Common packages** - `agent-builder-common`: Added `TodosStep`, `isTodosStep`, `TodosStepData` (with `carried_over` flag) to the `ConversationRoundStep` union; exported todo tool constants - `agent-builder-server`: Added `todos: boolean` to `ExperimentalFeatures`; new `TodoStateManager` (in-memory per-execution, initialized from conversation state); exported from runner index **Server / execution** - New built-in tool `server/services/tools/builtin/todo/todo_write.ts` — replaces the full todo list on each call - `select_tools.ts`: todo tool gated behind `experimentalFeatures.todos` - `run_chat_agent.ts`: captures `initialTodos` before the round starts, passes to `addRoundCompleteEvent` - `add_round_complete_event.ts`: uses `initialTodos` as fallback `TodosStep` with `carried_over: true` when agent never called `TodoWrite`; helper `carriedOverTodos` skips carryover when all todos are done - `run_agent.ts` / `runner.ts`: wires `todos` experimental flag and `todoStateManager` - `research_agent.ts`: includes `getTodoInstructions()` prompt section when `experimentalFeatures.todos` is enabled - New `prompts/utils/todos.ts`: opencode-inspired task management prompt explaining when/how to use the tool, including the full-list replacement semantic **Frontend** - New `todos_step_display.tsx`: collapsed header-only view when `carried_over`, full expanded list with `expandIn` keyframe animation on mount when not carried over - `round_layout.tsx`: renders `TodosStepDisplay` per-round between the thinking section and the response - `use_conversation_actions.ts`: `addOptimisticRound` seeds the new round with carried-over todos; new `addOrUpdateTodosStep` action flips `carried_over → false` triggering the expand animation - `use_subscribe_to_chat_events.ts`: wired to call `addOrUpdateTodosStep` on todo events - `types.ts` (conversation client): `TodosStep` added to `PersistentConversationRoundStep` for ES serialization **Tests** - `chat_message_text.test.tsx`: added `addOrUpdateTodosStep: jest.fn()` to conversation actions mock - `test_utils/runner.ts`: added `todos: false` to `experimentalFeatures` mock ## Test plan - [ ] Enable `agentBuilder:experimentalFeatures` in Kibana advanced settings - [ ] Start a new conversation with a multi-step task; confirm agent creates todos and they appear in the round. e.g. "Can you please search for security alerts with these keywords (using security.alerts), one by one: Virus, malware, macOS. Create a to-do list for each." - [ ] Start a second round with incomplete todos (by sending something like "Add windows to the list"); confirm they carry over collapsed ("To-dos N"), then expand when agent calls `TodoWrite` - [ ] Start a new round that does not modify todos (e.g. say "Thanks") - confirm todos were carried over - [ ] Ask the agent to "Complete the todos sequentially" - observe all todos being marked as completed - [ ] Say "Thanks" to the agent - observer todos are not carried over because they are all complete - [ ] Refresh the page; confirm the todo state and collapsed/expanded display is preserved from persisted round data - [ ] Disable the experimental setting; confirm the todo tool is not available, and the todo prompt section is not part of the prompt anymore (check inside of Langsmith or other observability tooling) --------- Co-authored-by: kibanamachine <42973632+kibanamachine@users.noreply.github.com> Co-authored-by: Elastic Machine <elasticmachine@users.noreply.github.com>

…c#265578) ## Summary Part of elastic/search-team#13832 <img width="960" height="702" alt="image" src="https://github.com/user-attachments/assets/2995f47e-7f33-463f-a426-2c07c337c811" /> https://github.com/user-attachments/assets/13cd178c-2f8c-4d30-afea-75ef460a2125 - Adds a new `platform_core_todo_write` built-in tool that lets the research agent plan and track work during a conversation round - Introduces `TodosStep` as a member of `ConversationRoundStep` so each round permanently records its todo snapshot — visible when scrolling through conversation history - Incomplete todos carry over automatically from the previous round; if all todos completed, nothing carries over - Frontend collapses carried-over todos to a "To-dos N" header and expands with a CSS animation when the agent calls `TodoWrite` during the round - Tool and system prompt are gated behind the existing `agentBuilder:experimentalFeatures` UI setting ## Changes **Common packages** - `agent-builder-common`: Added `TodosStep`, `isTodosStep`, `TodosStepData` (with `carried_over` flag) to the `ConversationRoundStep` union; exported todo tool constants - `agent-builder-server`: Added `todos: boolean` to `ExperimentalFeatures`; new `TodoStateManager` (in-memory per-execution, initialized from conversation state); exported from runner index **Server / execution** - New built-in tool `server/services/tools/builtin/todo/todo_write.ts` — replaces the full todo list on each call - `select_tools.ts`: todo tool gated behind `experimentalFeatures.todos` - `run_chat_agent.ts`: captures `initialTodos` before the round starts, passes to `addRoundCompleteEvent` - `add_round_complete_event.ts`: uses `initialTodos` as fallback `TodosStep` with `carried_over: true` when agent never called `TodoWrite`; helper `carriedOverTodos` skips carryover when all todos are done - `run_agent.ts` / `runner.ts`: wires `todos` experimental flag and `todoStateManager` - `research_agent.ts`: includes `getTodoInstructions()` prompt section when `experimentalFeatures.todos` is enabled - New `prompts/utils/todos.ts`: opencode-inspired task management prompt explaining when/how to use the tool, including the full-list replacement semantic **Frontend** - New `todos_step_display.tsx`: collapsed header-only view when `carried_over`, full expanded list with `expandIn` keyframe animation on mount when not carried over - `round_layout.tsx`: renders `TodosStepDisplay` per-round between the thinking section and the response - `use_conversation_actions.ts`: `addOptimisticRound` seeds the new round with carried-over todos; new `addOrUpdateTodosStep` action flips `carried_over → false` triggering the expand animation - `use_subscribe_to_chat_events.ts`: wired to call `addOrUpdateTodosStep` on todo events - `types.ts` (conversation client): `TodosStep` added to `PersistentConversationRoundStep` for ES serialization **Tests** - `chat_message_text.test.tsx`: added `addOrUpdateTodosStep: jest.fn()` to conversation actions mock - `test_utils/runner.ts`: added `todos: false` to `experimentalFeatures` mock ## Test plan - [ ] Enable `agentBuilder:experimentalFeatures` in Kibana advanced settings - [ ] Start a new conversation with a multi-step task; confirm agent creates todos and they appear in the round. e.g. "Can you please search for security alerts with these keywords (using security.alerts), one by one: Virus, malware, macOS. Create a to-do list for each." - [ ] Start a second round with incomplete todos (by sending something like "Add windows to the list"); confirm they carry over collapsed ("To-dos N"), then expand when agent calls `TodoWrite` - [ ] Start a new round that does not modify todos (e.g. say "Thanks") - confirm todos were carried over - [ ] Ask the agent to "Complete the todos sequentially" - observe all todos being marked as completed - [ ] Say "Thanks" to the agent - observer todos are not carried over because they are all complete - [ ] Refresh the page; confirm the todo state and collapsed/expanded display is preserved from persisted round data - [ ] Disable the experimental setting; confirm the todo tool is not available, and the todo prompt section is not part of the prompt anymore (check inside of Langsmith or other observability tooling) --------- Co-authored-by: kibanamachine <42973632+kibanamachine@users.noreply.github.com> Co-authored-by: Elastic Machine <elasticmachine@users.noreply.github.com>

KDKHD added 6 commits April 24, 2026 17:49

[AgentBuilder] Update todo prompt wording and carry over dead code files

c5e7fa6

- Adds explicit rule that marking a task in_progress and completing the previous can happen in the same TodoWrite call - Carries over todo_panel.tsx and use_todo_list.ts (unused legacy files, kept for reference)

[AgentBuilder] Fix todos.ts prompt update not staged in previous commit

ae85ffb

[AgentBuilder] Add todo UI hint to answer agent prompt

b4f21a4

[AgentBuilder] Apply linter fixes to answer_agent and todos prompts

08cda1e

KDKHD marked this pull request as ready for review April 30, 2026 15:00

KDKHD requested a review from a team as a code owner April 30, 2026 15:00

infra-vault-gh-plugin-prod Bot deleted a comment from elasticmachine Apr 30, 2026

chrisbmar reviewed May 6, 2026

View reviewed changes

Comment thread ...form/plugins/shared/agent_builder/server/services/execution/run_agent/prompts/utils/todos.ts Outdated

chrisbmar changed the title ~~[AgentBuilder] Add todo list tool and per-round TodosStep UI~~ [Agent Builder] Add todo list tool and per-round TodosStep UI May 6, 2026

pgayvallet reviewed May 6, 2026

View reviewed changes

@chrisbmar feedback

cb7db01

KDKHD force-pushed the feature-agent-builder-todo-list branch from cce0031 to cb7db01 Compare May 6, 2026 11:51

kibanamachine and others added 4 commits May 6, 2026 12:17

Changes from node scripts/eslint_all_files --no-cache --fix

188da60

@pgayvallet feedback

701e733

Merge branch 'feature-agent-builder-todo-list' of github.com:KDKHD/ki…

e043180

…bana into feature-agent-builder-todo-list

Changes from node scripts/eslint_all_files --no-cache --fix

ab059be

KDKHD added release_note:skip Skip the PR/issue when compiling release notes backport:skip This PR does not require backporting labels May 6, 2026

KDKHD added 2 commits May 7, 2026 13:38

Merge remote-tracking branch 'upstream/main' into feature-agent-build…

579b2dd

…er-todo-list # Conflicts: # x-pack/platform/packages/shared/agent-builder/agent-builder-common/index.ts # x-pack/platform/plugins/shared/agent_builder/server/services/execution/run_agent/prompts/answer_agent.ts

Merge branch 'feature-agent-builder-todo-list' of github.com:KDKHD/ki…

70d9415

…bana into feature-agent-builder-todo-list

KDKHD requested review from chrisbmar and pgayvallet May 8, 2026 13:31

pgayvallet approved these changes May 11, 2026

View reviewed changes

Merge branch 'main' into feature-agent-builder-todo-list

7b2cc55

chrisbmar approved these changes May 11, 2026

View reviewed changes

KDKHD merged commit 4f6b28f into elastic:main May 11, 2026
32 checks passed

kibanamachine added the v9.5.0 label May 11, 2026


		export const getTodoInstructions = (): string => `## TASK MANAGEMENT

		You have access to the TodoWrite tool to plan and track your work. Use it frequently so the user can see what you are doing and how you are progressing.

		const carriedTodos = carriedOverTodos(initialTodos);
		const todosForStep = lastTodosData ?? carriedTodos;

Conversation

KDKHD commented Apr 24, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Changes

Test plan

Uh oh!

chrisbmar left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

pgayvallet left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

pgayvallet May 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

KDKHD commented May 7, 2026

Uh oh!

pgayvallet left a comment

Choose a reason for hiding this comment

Uh oh!

KDKHD commented May 11, 2026

Uh oh!

chrisbmar left a comment

Choose a reason for hiding this comment

Uh oh!

kibanamachine commented May 11, 2026

💛 Build succeeded, but was flaky

Failed CI Steps

Test Failures

Metrics [docs]

Module Count

Async chunks

API count

History

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

KDKHD commented Apr 24, 2026 •

edited

Loading

chrisbmar left a comment •

edited

Loading

pgayvallet May 6, 2026 •

edited

Loading