Planning and Task Breakdown skill for Claude Code: turn a spec into ordered tasks

Addy Osmani's skill that makes Claude split a spec into small, ordered tasks with acceptance criteria, for work that feels too big to start.

Installs: 68.7K; on skills.sh
Stars: 112.6K; on GitHub
Updated: Nov 2026; last repo commit

When you have a spec or clear requirements, Claude first goes into read-only plan mode. It maps what depends on what, slices the work into vertical features that work end to end, and writes each task with acceptance criteria, verification and files likely touched. It adds checkpoints after every two or three tasks.

It saves the plan to tasks/plan.md and the task list to tasks/todo.md, or to an external tracker if your project designates one. It will not overwrite an old plan that still has unchecked tasks without asking. It bundles no scripts.

Good for

Developers with a spec who do not know where to start
Leads who need to explain scope to a human
People running several agents or sessions in parallel

Install planning-and-task-breakdown

Claude Code

Copy this prompt.

Install the planning-and-task-breakdown skill in Claude Code. Run this command:

/plugin marketplace add addyosmani/agent-skills
/plugin install agent-skills@addy-agent-skills

Then check that it loaded and tell me how to use it.

2
Paste it into Claude Code. It shows the change and asks before running it.

The skill links to a shared references/definition-of-done.md in the pack, and the /plan and /build commands come with the pack.

From the project README. Source.

What's in the skill

When it loads

Breaks work into ordered tasks. Use when you have a spec or clear requirements and need to break work into implementable tasks. Use when a task feels too large to start, when you need to estimate scope, or when parallel work is possible.

Sections in SKILL.md

Overview
When to Use
The Planning Process
Step 1: Enter Plan Mode
Step 2: Identify the Dependency Graph
Step 3: Slice Vertically
Step 4: Write Tasks
Task [N]: [Short descriptive title]
Step 5: Order and Checkpoint
Checkpoint: After Tasks 1-3
Task Sizing Guidelines
Output Files
Task List Target
Plan Document Template
Overview
Architecture Decisions

Key facts

Author: addyosmani (agent-skills pack)
Repository: addyosmani/agent-skills
Installs on skills.sh: 68,687 (2 Nov 2026)
GitHub stars (repo): 112,555 (2 Nov 2026)
License: MIT
Last commit (repo): 1 Nov 2026
What installs: SKILL.md only (1 file); bundles no scripts. Its footer points to a shared references/definition-of-done.md in the pack
Works in: Claude Code, Cursor, Codex, Copilot, Cline and more through the skills CLI; native Claude Code and Codex plugins
Needs: The wider agent-skills pack for the shared definition-of-done reference it links to
skills.sh audits: Gen Agent Trust Hub pass, Socket pass, Snyk pass

What it includes

Read the spec and code, write no code
Claude works read-only, maps patterns and risks, and outputs a plan document, not code.
Read · Step 1: Enter plan mode
Map what depends on what, build foundations first
Implementation order follows the graph bottom-up.
Read · Step 2: Dependency graph
One complete feature path per task
Claude prefers tasks like "user can log in" over "build all the database".
Read · Step 3: Slice vertically
Description, acceptance, verification, dependencies, files, scope
Each task uses one template, including a size from small to large.
Try: “Break the approved spec in SPEC.md into tasks.”
Write · Step 4: Write tasks
Checkpoints after every 2 to 3 tasks, risky tasks early
Claude adds checkpoints such as tests pass, app builds, core flow works, and review with a human.
Write · Step 5: Order and checkpoint
XS to XL; XL must be split
A task touching 8 or more files is too large and gets broken down.
Read · Task sizing
tasks/plan.md and tasks/todo.md, or an external tracker
Claude checks for an existing plan with unchecked tasks first and asks before replacing it. With a tracker such as GitHub Issues, Jira or Linear it creates one item per task.
Write · Output files

From the official docs

Permissions

Low risk

The skill writes plan files into your repo, or creates items in your issue tracker if your project names one. Check which one before you let it run.

What it can do

Read your data
Create and change things
Delete things
Read-only mode

Watch out for

It writes plan files, or creates tracker items
Plans go to tasks/plan.md and tasks/todo.md unless a tracker is designated. The skill says to ask before replacing a plan with unchecked tasks, and never to bulk-close another plan's tracker items.

Security advisories

No published security advisories as of Nov 2, 2026.

Technical details

Turning on read-only	The skill defines no read-only switch.
Talks to	Your issue tracker, if you designate one
Scopes	Claude writes tasks/plan.md and tasks/todo.md With a designated tracker, Claude creates one tracker item per task

Questions

What does the Planning and Task Breakdown skill do?

It makes Claude turn a spec into a plan and a checklist of small tasks. Each task has acceptance criteria, a way to verify it, dependencies and an expected size.

How do I install it?

Use the install box at the top of this page. It installs the whole agent-skills pack, which also gives you the /plan command and the other lifecycle skills. A per-skill install misses the pack's shared references folder.

Where does the plan go?

By default into tasks/plan.md (the plan) and tasks/todo.md (the checklist) in your repo. If your CLAUDE.md or AGENTS.md names an issue tracker, Claude creates tracker items instead and keeps only an index in the plan.

Will it overwrite my existing plan?

Not without asking. If tasks/plan.md or tasks/todo.md still has unchecked tasks for different work, the skill tells Claude to stop and let you decide. It updates in place only when you are revising the same plan.

Planning and Task Breakdown vs spec-driven-development?

Spec-driven-development writes the spec. This skill takes the approved spec and breaks it into tasks. Spec-driven-development calls this skill the canonical source for planning.

Does it work in Cursor, Codex or Copilot?

Yes. The README documents a Codex plugin, the skills CLI for 70+ agents including Cursor, and manual setup for Cursor and Copilot. The slash commands (/spec, /plan, /build and so on) are Claude Code commands and do not appear in Copilot, where you invoke skills by name. The README does not describe a Claude app install.

Is it safe to install?

It is low risk. It bundles no scripts. It writes plan files, or creates tracker items if you use a tracker. Check git status after a run, and check which tracker project it targets before you let it create items.

Information

updated Nov 2, 2026

Repository: addyosmani/agent-skills
Folder: skills/planning-and-task-breakdown

addyosmani/agent-skills skills.sh

Sources

Docs, issue trackers and discussions this page is built from.

Show all 3

addyosmani/agent-skills README · github.com
planning-and-task-breakdown SKILL.md · github.com
github.com · github.com