Skip to content
Get packs

Guide · Agent skill

How to use the Daily Task Manager agent skill

Updated · written with AI from the project's own documentation and checked against our catalogue data

The Daily Task Manager is an agent skill that keeps a running to-do list for you inside a gbrain knowledge store, in a page called ops/tasks.md. You talk to your agent in plain language to add, complete, defer or review tasks, and it records each change with a stable ID, a priority and an audit trail. This guide explains what an agent skill is, how to install this one, how to trigger it, and what it needs to work.

Daily Task Managergarrytan/gbrain · MIT

Quick start

  1. 1Download and unzip. Download the ZIP from this skill's page and unzip it to get the daily-task-manager folder.
  2. 2Place or upload the skill. Copy the folder to ~/.claude/skills/ or a project's .claude/skills/, or upload the ZIP under Settings → Capabilities → Skills in the Claude apps.
  3. 3Confirm your gbrain tools. Make sure your environment provides search, get_page, put_page and add_timeline_entry.
  4. 4Add your first task. Ask the agent to add a task and give a priority. The first run creates ops/tasks.md if it does not exist.
  5. 5Review your list. Ask what your tasks are, or what is on your plate today, to see active tasks grouped by priority.
  6. 6Complete or defer. Say you finished a task to complete it, or give a target date and reason to defer it.

What an agent skill is and what this one does

An agent skill is a folder containing a SKILL.md file that tells an AI agent how to handle a specific kind of request. This skill is named daily-task-manager and is designed for task lifecycle management: adding tasks, completing them, deferring them, removing them and reviewing what is open.

Tasks live in one brain page, ops/tasks.md, inside a gbrain knowledge store. Each task gets a stable ID in the form t-YYYYMMDD-NN, a priority from P0 (urgent) to P3 (backlog), and optional due or defer dates. Completed tasks are moved to a Completed section with the completion date, and deferred tasks carry a target date and a reason.

  • P0: urgent
  • P1: today
  • P2: this week
  • P3: backlog
  • Lifecycle: add, then in-progress, then complete, defer or remove

Install the skill from a ZIP or the Claude apps

Download the ZIP from the skill's page on this site and unzip it. For a personal install, place the resulting daily-task-manager folder in ~/.claude/skills/. For a single project, place it in that project's .claude/skills/ folder instead.

If you use the Claude apps, you can upload the skill instead. Open Settings, go to Capabilities, then Skills, and upload the ZIP there.

  • Personal install: ~/.claude/skills/
  • Project install: .claude/skills/ inside the project
  • Claude apps: Settings → Capabilities → Skills

Check that your gbrain environment is ready

The skill does not work on its own. It depends on a gbrain brain environment that provides the search, get_page, put_page and add_timeline_entry tools. Without these, the agent has no way to read or save ops/tasks.md.

The skill file itself declares no API keys. It does not run scripts and does not use the network.

  • Required tools: search, get_page, put_page, add_timeline_entry
  • No API keys declared in SKILL.md

Trigger the skill with natural requests

You do not need special syntax. The skill maps your intent to an action. Phrases like add a task, remind me to, or put X on my list become an add. Done with X, finished X or completed X mark a task complete. Push X, defer X or move X to next week defer it. Delete X, remove X or kill task X remove it. What are my tasks, task list or what's on my plate today trigger a review.

Remove only happens on explicit delete words. The skill never infers a removal from context.

  • "Add task: renew passport, P1, due 2026-02-10"
  • "Done with the passport renewal"
  • "Push the budget review to next week, reason: waiting on numbers"
  • "What's on my plate today?"
  • "Delete the old dentist task"

How the agent handles missing details and ambiguity

If you leave out something required, the agent asks one concise clarifying question before changing anything. It never invents priorities, due dates or defer reasons. If you add a task without a priority, it uses P3 and tells you it did so.

If your wording matches more than one active task, the agent lists the candidates with their IDs and asks which one you mean. It does not guess the closest match. Deferring requires both a target date and a reason, and removing asks for confirmation unless your message already says so.

  • Missing priority on add: defaults to P3 and says so
  • Ambiguous match: lists candidates and asks
  • Defer needs a target date and a reason
  • Remove needs explicit confirmation

Get a machine-readable list for agents

For tools and sub-agents, the skill can return results in a structured form. Every action reports a result with the action, task_id, status, priority, date, page and a saved flag. When invoked with the trigger task list json, the skill returns a JSON array of task objects with id, description, priority, due and status.

Because the task cycle is read-modify-write, run task changes one at a time.

  • Result fields: action, task_id, status, priority, date, page, saved
  • Statuses: ok, not_found, ambiguous, needs_confirmation
  • JSON trigger: task list json

What people use it for

  • Keeping a running personal to-do list stored alongside your notes in gbrain.
  • Asking the agent each morning what is urgent or due today, using the today filter that shows P0 and P1.
  • Deferring a task with a target date and reason so it moves out of the active list.
  • Having a sub-agent read the task list as JSON and chain it into another workflow.

Frequently asked questions

Is the Daily Task Manager skill free?

The source is published under the MIT licence. The skill file declares no API keys, but it requires a gbrain brain environment to run.

What do I need to use this skill?

You need a gbrain brain environment that provides the search, get_page, put_page and add_timeline_entry tools. The skill does not work as a standalone task app.

Does it send reminders or notifications?

No. The skill only records tasks when you ask. It does not send notifications, so do not rely on it for critical deadlines.

Can several people edit the same task list at once?

Not safely. The skill's own documentation warns that simultaneous writes can overwrite each other and create duplicate IDs, so task edits must be run one at a time.

Does the skill run scripts or use the network?

According to its declaration, it runs no scripts and uses no network. It works through the gbrain page tools.

How does it compare to a standalone to-do app?

It keeps tasks as a page inside your gbrain brain, so they sit alongside your notes and can be read by agents. If you do not use gbrain, a standalone task app is the better fit.

New agent skills, repos and AI tools — every week

The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.

Our verdict

This skill suits people already using gbrain who want an AI assistant to maintain a structured, prioritised task list next to their notes, and developers who need JSON task output for agent chains. Skip it if you do not run gbrain, if several people must edit one list at the same time, or if you need reminders for critical deadlines.

Download the skill