How to use the Sprint Planning Assistant agent skill

What the sprint-plan skill makes the agent do

The skill is a single file called SKILL.md. When it is installed and you ask the agent to plan a sprint, the agent follows the steps written in that file instead of improvising its own approach.

It estimates how much work the team can handle, accounting for people's availability and a 15-20% buffer for unexpected work, bugs and tech debt. It then selects stories from the prioritized backlog, checks that each one is ready, orders dependent stories, lists risks with mitigations and writes a summary with a single sprint goal.

• Capacity estimation based on team members, availability (PTO, meetings, on-call) and historical velocity from the last 3 sprints
• Story selection from the highest-priority backlog items, stopping when capacity is reached
• Dependency mapping, including external teams, and identification of the critical path
• Risk identification for uncertain stories, slipping external dependencies and knowledge concentration, each with a suggested mitigation
• A sprint plan summary and one-sentence sprint goal

Install the sprint-plan skill

Download the ZIP file for the skill from this directory page. Then choose one of the two installation routes below.

For a personal setup available across your projects, unzip the download into ~/.claude/skills/. For a single project, unzip it into the .claude/skills/ folder inside that project. In both cases the skill folder should contain SKILL.md.

If you use the Claude apps instead, upload the ZIP under Settings → Capabilities → Skills.

• Personal install: ~/.claude/skills/
• Project install: .claude/skills/ inside your project folder
• Claude apps: Settings → Capabilities → Skills, then upload the ZIP

How to trigger the skill with a natural request

You do not need a special command. Describe the planning task in plain language, naming the team or project. The skill's context uses that name as the subject of the plan, so be specific about which team the sprint is for.

If you have files, attach or point the agent to them before asking. The skill tells the agent to read backlogs, velocity data, team rosters or previous sprint reports first, so the more relevant material you provide, the more grounded the draft will be.

• Plan the next sprint for the checkout team, using the attached backlog and velocity history.
• Estimate capacity for our two-week sprint. Two engineers are on PTO on Wednesday and Thursday.
• Which stories from the prioritized backlog fit into the next sprint, and what depends on the payments team?
• Balance our sprint scope against the velocity from the last three sprints and list the risks.

Read the sprint plan the agent produces

The agent writes its output in a fixed structure: sprint goal, duration, team capacity, committed stories with points and owners, buffer, then a list of risks with mitigations. Because the format is consistent, you can compare plans from one sprint to the next.

Treat the result as a draft. The estimates and risks come from the text you provided, not from your issue tracker or real project data, so check each story against your team's own knowledge before the planning meeting.

• Sprint Goal: one sentence describing what success looks like
• Team Capacity and Committed Stories, shown in story points
• Buffer: the capacity left over after commitments
• Stories: title, points, owner and dependencies for each item
• Risks: each risk paired with a mitigation

Safety notes: scripts and network access

The skill is instructions only. The listing states that it runs no scripts and uses no network access, and it has no listed requirements. Its only file is SKILL.md.

The skill's output is text generated from the material you supply. Review it like any draft written by a colleague, and keep sensitive backlog or roster details out of the prompt if your organisation restricts what you share with an AI assistant.

• Runs scripts: no
• Uses network: no
• Requires: nothing listed
• Files included: SKILL.md

Tips for better sprint plans

The skill depends on inputs it does not supply itself. If you do not give it a prioritized backlog or velocity history, the agent may guess, so provide those first. Teams without a backlog or velocity history will get the weakest results.

Use the draft to start the conversation, not to end it. Share it with the team, adjust the capacity buffer for your situation and confirm the risks and external dependencies with their owners.

• Provide a prioritized backlog so stories are chosen in the right order
• Include velocity from the last three sprints, as the skill's velocity step uses that window
• Attach a roster with availability, including PTO, meetings and on-call duty
• Ask the agent to flag stories that need refinement before they are committed
• Review the critical path and external owners with the team before finalising