Planning and Task Breakdown
Splits a spec into small, ordered, verifiable tasks before any code is written.
- Repository
- addyosmani/agent-skills
- Stars
- 112,668
- Licence
- MIT
- Files
- 1 file
What it does
When an agent is given a spec or a large vague request, this skill makes it plan first instead of coding. It reads the relevant code without changing it, maps which parts depend on which, and splits the work into small tasks with clear acceptance criteria, verification steps, dependencies and size estimates. It saves the plan to a markdown file (tasks/plan.md) and a checklist (tasks/todo.md), or to an external issue tracker if the project specifies one, and adds review checkpoints between groups of tasks.
Great for
- Developers handing a feature spec to an AI coding agent who want reliable, reviewable steps
- Teams that need to estimate scope or split work across several agents or sessions
- Anyone whose agent keeps producing tangled, half-finished changes on large tasks
Skip it if
- Single-file or obviously scoped changes, which the skill itself says it should not be used for, because the planning overhead adds little value
- Non-developers looking for general project planning outside a codebase, since the templates assume code, tests and build commands
- Projects where an existing plan file has unchecked tasks: the skill tells the agent to stop and ask rather than overwrite, so it may pause the session
Try asking
- “Here is our spec for user authentication. Break it into ordered implementation tasks with acceptance criteria.”
- “This task feels too big to start. Plan it into small vertical slices I can review one at a time.”
- “Estimate the scope of adding search with pagination and tell me which tasks could run in parallel.”
- “Create a plan for this feature and record the tasks as Linear issues, since our AGENTS.md says we use Linear.”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 1 original files, the MIT licence and an attribution note.
planning-and-task-breakdown.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip planning-and-task-breakdown.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip planning-and-task-breakdown.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Give the agent a clear spec or requirements first; the skill works best when there is something concrete to break down.
- Name your issue tracker in CLAUDE.md or AGENTS.md if you use one, so tasks go there instead of tasks/todo.md.
- Review the plan and checkpoints before letting the agent start building, since the skill is designed to pause for human review.
What's inside
- SKILL.md
Redistributed unmodified from addyosmani/agent-skills under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.