Guide · Agent skill
How to use the Chat TODO plan agent skill (todos)
Updated · written with AI from the project's own documentation and checked against our catalogue data
The todos skill teaches an AI agent to keep a visible plan for multi-step work in the current chat. It is written for the JetBrains ThinkRail project's Todo panel, so the agent can organize a task into groups and steps that you can follow at a glance. After this guide you will know what the skill does, how to install it, how to ask for a plan in natural language, and what it needs to run.
Chat TODO planJetBrains/thinkrail · Apache-2.0 Download the skill Quick start
- 1Download the ZIP. Get the todos skill ZIP from this page and unzip it.
- 2Place the folder. Move the todos folder into ~/.claude/skills/ for personal use or into your project's .claude/skills/ folder.
- 3Or upload in the Claude apps. In the Claude apps, open Settings, go to Capabilities, then Skills, and upload the skill.
- 4Confirm the Todo tools. Make sure your environment is the Todo panel of JetBrains ThinkRail and provides the todo_* tools the skill needs.
- 5Ask for a multi-step task. Describe a task with at least three steps or ask explicitly for a plan.
- 6Watch the Todo panel. Follow statuses as the agent moves steps from pending to in progress to done, and add or remove tasks if needed.
- 7Review the summary. At the end, read the summary and verification for each completed code step before accepting the work.
What an agent skill is and what the todos skill makes the agent do
An agent skill is a folder containing a SKILL.md file with instructions that an AI agent reads when a request matches the skill's description. The todos skill is one such file. Its description says it applies when you ask for a shared plan, when a task needs at least three substantive execution steps, or when a loose TODO is pending.
Once active, the agent organizes multi-step work into a plan. Each task you give becomes a group whose title describes the outcome, and the steps inside are items with a title, a status (pending, in progress, or done) and an optional note. The plan lives with the chat session, not in the code repository, and you see it in the Todo panel.
- One group per task, titled by the outcome (for example "Fix login redirect").
- Steps are items with statuses: pending, in_progress, done.
- The agent works through steps in order and re-reads the list so it stays in sync with your edits.
- At the end it summarizes what changed and how it was verified.
Install the todos skill (ZIP download or Claude apps)
Download the ZIP from this page and unzip it. For a personal setup, place the unzipped skill 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 skill there.
- Personal install: ~/.claude/skills/todos
- Project install: .claude/skills/todos inside your project
- Claude apps: Settings → Capabilities → Skills
~/.claude/skills/Check that the requirements are met
The skill depends on a todo tool environment that provides todo_write, todo_list, todo_add, todo_promote and todo_remove. The source identifies this as the Todo panel of the JetBrains ThinkRail project. If those tools are not available, the skill cannot do its job, and the source does not say what happens in that case.
The skill runs no scripts and uses no network access, according to its listing.
- Required tools: todo_write, todo_list, todo_add, todo_promote, todo_remove
- Required environment: the Todo panel of JetBrains ThinkRail
- Runs scripts: no
- Uses network: no
Trigger the skill with natural requests
You do not need a special command. Ask for what you want in plain language. Explicitly asking for a plan always triggers the skill. A larger task that needs at least three substantive steps will also trigger it.
The skill is designed to stay out of the way for quick work. Single answers, simple checks, and one- or two-step edits do not get a plan unless you ask for one. If you have already added a loose item to the list, the agent may take it up as its own group.
- "Make a shared plan for migrating the settings page, then start on it."
- "Refactor the export module into three steps and track them in the plan."
- "Add a task to the list: update the README install section."
- Quick questions such as "What does this function return?" do not get a plan.
Add, remove and reorder tasks while the agent works
The plan is shared. You can add tasks and remove ones you no longer want while the agent is working. The agent re-reads the list before choosing its next item, so your edits take effect. If a step the agent planned has disappeared, you dropped it, and the agent will skip it rather than re-add it.
New loose items appear at the end of the plan. The agent finishes the task it is on before picking up a loose item, so a freshly added item does not interrupt work in progress.
- Add a loose item from the UI and it queues after the current work.
- Remove an item you dropped and the agent will not bring it back.
- Steps the agent discovers mid-task are inserted under the current step.
Read the summaries and verification
When a step changes code, the agent marks it done with a summary and a verification. The summary explains what changed and why, including decisions that the diff does not show. The verification names the check that was run and its result, or says it was not verified.
Read these alongside the code diff. The summary and verification are the record you use to review the session.
- Summary: what changed and why, in structured Markdown.
- Verification: the check and its result, such as a test command and its outcome, or "not verified".
What people use it for
- Multi-step feature work where you want to see progress steps without reading the whole chat.
- Splitting a large refactor into commit-sized, checkable steps.
- Keeping a shared list of follow-up tasks you add while the agent is busy.
- Producing a clear record of what was changed and verified in a session.
Frequently asked questions
How do I install the todos agent skill?
Download the ZIP from this page and unzip it. Put the todos folder in ~/.claude/skills/ for personal use, or in a project's .claude/skills/ folder for one project. In the Claude apps you can instead upload it under Settings → Capabilities → Skills.
What are the requirements for the todos skill?
It needs a todo tool environment providing todo_write, todo_list, todo_add, todo_promote and todo_remove, which the source identifies as the Todo panel of the JetBrains ThinkRail project. It runs no scripts and uses no network. The source does not say what happens without the Todo tools.
Is the todos skill free?
The source lists the licence as Apache-2.0. It does not mention pricing, so check the repository for any cost details.
Does the todos skill run code or access the network?
According to its listing, it runs no scripts and uses no network. It works through the todo_* tools and the plan is stored with the chat session, not committed to the repository.
Why doesn't the agent make a plan for my simple question?
The skill is designed to stay out of the way for one-shot answers, checks, and one- or two-step edits unless you explicitly ask for a plan. Add the word plan to your request if you want one for a small task.
What should I do if the Todo panel is not available?
The source says the skill depends on the Todo panel and todo_* tools, but it does not say what happens without them. Use an environment that provides those tools, or consider a different skill for planning.
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
Use the todos skill if you run multi-step tasks with an agent in a JetBrains ThinkRail environment and want a live, shared plan with a clear review trail. Pick something else if your setup has no Todo panel or todo_* tools, or if you mostly ask quick questions and small edits, since the skill is built to stay quiet for those.
Download the skill