agentleFS
cognitedatacognitedata/builder-skills/skills/integrate-todo-list/SKILL.md

integrate-todo-list

Skill6 starsChanged 9 days agoApache-2.0View on GitHub

TodoWrite panel for an already-approved in-app useAtlasChat UI. New AI features use integrate-fusion-agent. Triggers: TodoWrite, todo panel — only when useAtlasChat already exists.

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 2 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Add a TodoWrite tool and panel so the in-app Atlas agent can show task progress.

    line 9
  2. Distinctive:

    Before writing anything, read:

    line 17
  3. Very specific:

    Read each file from <skill-dir>/code/ and write it into src/todo/ with the same filename:

    line 30
  4. Very specific:

    All files use relative imports (./types, ./TodoContext, etc.) — no changes needed.

    line 43
  5. Very specific:

    Run the app's type-check command (typically pnpm tsc --noEmit) and confirm there are no errors.

    line 145
  6. Distinctive:

    If the project has tests, run them to confirm nothing regressed.

    line 146
  7. Very specific:

    Update task status in real-time (pending → in_progress → completed)

    line 154

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Integrate Todo List

Add a TodoWrite tool and panel so the in-app Atlas agent can show task progress.

Skip if the app uses the Atlas / EOS sidebar (integrate-fusion-agent). Prerequisite: useAtlasChat + src/atlas-agent/ + @sinclair/typebox from integrate-atlas-chat.


Step 1 — Read the app

Before writing anything, read:

  • package.json — confirm @tabler/icons-react is installed; if not, install it with the app's package manager
  • src/App.tsx — find where to add TodoProvider
  • The file that calls useAtlasChat (likely src/chat/useChatViewModel.ts or src/App.tsx) — this is where the tool gets wired
  • The chat view component that renders messages — this is where TodoPanel and TodoToolResultCard go

Step 2 — Create the src/todo/ module

Find the skill directory by running find . -path "*/.agents/skills/integrate-todo-list/code" -type d from the project root.

Read each file from <skill-dir>/code/ and write it into src/todo/ with the same filename:

FilePurpose
types.tsTodoItem and TodoList types
TodoContext.tsxReact context + TodoProvider
useTodoList.tsHook to read/write the todo list
todoWriteTool.tscreateTodoWriteTool factory — AtlasTool with full CDF task-decomposition guidance
useTodoWriteTool.tsHook that memoizes the tool with current state access
TodoPanel.tsxCard UI: progress bar + task rows
TodoItemRow.tsxSingle row with animated status icons
TodoToolResultCard.tsxCompact summary card for tool call display

All files use relative imports (./types, ./TodoContext, etc.) — no changes needed.


Step 3 — Wrap the app in TodoProvider

In src/App.tsx (or the root component), wrap the existing tree with <TodoProvider>:

import { TodoProvider } from './todo/TodoContext'; // adjust path to match app conventions

function App() {
  return (
    <TodoProvider>
      {/* existing children */}
    </TodoProvider>
  );
}

Step 4 — Wire the tool into useAtlasChat

In the file that calls useAtlasChat, add the following. Adjust import paths to match the app's conventions.

import { useRef, useCallback } from 'react';
import { useTodoList } from './todo/useTodoList';
import { useTodoWriteTool } from './todo/useTodoWriteTool';

// Inside the hook/component:
const { todos, setTodos } = useTodoList();
const todoWriteTool = useTodoWriteTool();

// Keep a ref so getAppContext always reads fresh state without re-creating the callback.
const todosRef = useRef(todos);
todosRef.current = todos;

const getAppContext = useCallback(() => {
  const t = todosRef.current;
  if (t.length === 0) return undefined;
  const lines = t.map((item, i) => `${i + 1}. [${item.status}] ${item.content}`);
  return `Current todo list:\n${lines.join('\n')}`;
}, []);

// Add to useAtlasChat options:
const { messages, send, isStreaming, progress, error, reset, abort } = useAtlasChat({
  client: isLoading ? null : sdk,
  agentExternalId: AGENT_EXTERNAL_ID,
  tools: [todoWriteTool],   // add alongside any existing tools
  getAppContext,
});

// In the reset handler, clear the todo list:
const handleReset = useCallback(() => {
  reset();
  setTodos([]);
}, [reset, setTodos]);

// Expose todos in the return value so the view can render TodoPanel:
return { ..., todos };

Step 5 — Render TodoPanel in the chat view

In the component that renders the chat input area, add <TodoPanel> above the input field:

import { TodoPanel } from './todo/TodoPanel'; // adjust path

// In the render:
<TodoPanel todos={todos} />
<YourChatInput ... />

TodoPanel returns null when the list is empty, so it's safe to always render it.


Step 6 — Render TodoToolResultCard for tool call steps

In the component that renders per-message tool calls (typically a steps accordion or similar), branch on the tool name:

import { TodoToolResultCard } from './todo/TodoToolResultCard'; // adjust path

{toolCalls.map((tc, i) =>
  tc.name === 'TodoWrite' ? (
    <TodoToolResultCard key={i} toolCall={tc} />
  ) : (
    <YourDefaultToolCallCard key={i} toolCall={tc} />
  )
)}

Step 7 — Verify

Run the app's type-check command (typically pnpm tsc --noEmit) and confirm there are no errors. If the project has tests, run them to confirm nothing regressed.


Done

The agent can now use TodoWrite to create and track tasks. It will:

  • Show a task panel as soon as it starts multi-step work
  • Update task status in real-time (pending → in_progress → completed)
  • Clear the list automatically when all tasks are done
  • Inject the current task list into each prompt via getAppContext so it knows where it left off

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.