Guide · Agent skill
How to use the Sprint Plan Builder agent skill (sprint-plan)
Updated · written with AI from the project's own documentation and checked against our catalogue data
The sprint-plan skill is an agent skill for Claude Code that drafts a sprint plan for you. It reads the stories you have marked Ready under production/epics, the previous sprint's file and any design docs, then writes a first-draft plan with a goal, a capacity buffer, prioritized tasks, carryover and risks. This guide explains what an agent skill is, how to install this one, how to run it and what to check before you trust the output.
Quick start
- 1Confirm your backlog. Make sure stories exist under production/epics and that each one has a '> **Status**' line set to Ready for work you want to plan.
- 2Install the skill. Download the ZIP from this page and unzip it into ~/.claude/skills/ for all projects or into the project's .claude/skills/ folder, or upload it under Settings → Capabilities → Skills in the Claude apps.
- 3Start a new plan. Run /sprint-plan new in your project. The agent reads the Ready stories, the previous sprint file and any design docs.
- 4Set review depth if needed. Add --review full, lean or solo to change how many director gates run for that one plan.
- 5Review the draft. Check the sprint goal, the Must Have list, the 20% buffer and the carryover table. Answer any questions the agent asks before it writes the file.
- 6Check the goal prerequisites. Confirm the Goal Prerequisites table lists anything needed to show the goal working, since the skill reports these but does not add stories for them.
- 7Track the sprint. Use /sprint-plan status during the sprint to see where it stands.
What an agent skill is and what this one makes the agent do
An agent skill is a folder containing a SKILL.md file that tells a compatible agent how to perform a specific task. When the skill is invoked, the agent follows the instructions in that file, using the tools the file allows.
The sprint-plan skill is built for game projects using the Claude Code Game Studios framework. When you invoke it in new or update mode, the agent reads your story backlog, the previous sprint's file and relevant design docs, then drafts a sprint plan. The draft includes a sprint goal, a capacity section with a 20% buffer for unplanned work, Must Have, Should Have and Nice to Have task tables, a carryover table for unfinished work, a risk table and a Definition of Done checklist.
- Reads stories marked Ready under production/epics
- Reads the previous sprint file in production/sprints/ to understand velocity and carryover
- Scans design/gdd/ for context on the features those stories implement
- Checks production/milestones/ and production/risk-register/ if they exist
- Can run director-review gates before the plan is written
Install the sprint-plan skill from the ZIP or the Claude apps
Download the ZIP for this skill from this page. The skill is a single file, SKILL.md. Unzip it into one of two locations, depending on how you want to use it.
For a single project, put the skill in that project's .claude/skills/ folder. To make it available across your own projects, unzip it into ~/.claude/skills/. If you use the Claude apps instead, open Settings, go to Capabilities, then Skills, and upload the skill there.
- Personal use across projects: ~/.claude/skills/
- One project only: the .claude/skills/ folder inside that project
- Claude apps: Settings → Capabilities → Skills
~/.claude/skills/sprint-plan/SKILL.mdCheck your project has the required structure
The skill expects the Claude Code Game Studios repository structure. It looks for stories under production/epics, design documents under design/gdd and the hooks/yaml-helper.sh script. Each story file needs a Status line in the form '> **Status**' so the skill can find Ready stories.
If your project has no story backlog, the skill stops and tells you to run /create-stories first. It will not invent work items. At the standard or full rigor levels, run /create-epics before /create-stories.
- Story files at production/epics/[epic-slug]/story-NNN-[slug].md
- A Status line on each story, such as '> **Status**: Ready'
- Design docs in design/gdd/ (optional at minimal rigor, where design/game-brief.md is used instead)
Run the skill with a mode argument
Invoke the skill with one of three modes: new to draft a fresh sprint plan, update to revise an existing one, or status to see where the current sprint stands. You can add a review depth with --review full, --review lean or --review solo. The flag overrides the review mode for that run only.
In new mode the agent numbers the sprint as one above the highest sprint-NNN.md file in production/sprints/, or 001 if there are none. It then presents the plan for your review. It does not write the file until the gate phases have run and any revisions are made.
Phrase your request in plain terms, such as asking the agent to plan the next sprint from the Ready stories. The skill's documented entry points are the mode arguments shown here.
- new: draft a new sprint plan
- update: revise an existing plan
- status: report on the current sprint
- --review full | lean | solo: set review depth for one run
/sprint-plan new --review leanRead the output and the review gates
Review depth decides which checks run. In full mode a producer feasibility gate is spawned before the plan is written. In lean and solo modes that gate is skipped. The QA plan check runs in every mode. If you do not set a review mode, the skill follows modes.rigor: solo at minimal and lean at standard.
The skill never writes a review mode to project.yaml or production/review-mode.txt. To change depth for good, change modes.rigor, or pin a personal override in project.local.yaml.
- Carryover: unfinished stories appear only in the carryover table, never again as new work
- Story count per sprint follows story_granularity: 2 to 4 at coarse, 6 to 10 at balanced, 15 to 25 at fine
- A Ready backlog smaller than the range is planned whole, with no padding
/settings --local modes.review_mode=<full|lean|solo>Safety notes: scripts and network access
The skill runs a shell script. Its configuration step calls hooks/yaml-helper.sh with the resolve_config option to read project settings. It also lists the production/sprints/ folder. Its allowed tools include Read, Glob, Grep, Write, Edit, Agent and AskUserQuestion, along with those two shell commands.
The skill does not use the network. Review the SKILL.md file before installing if you want to confirm the exact tools it is allowed to use.
- Runs scripts: yes, the yaml-helper.sh config lookup and a directory listing
- Uses network: no
- It does not invent stories, so the plan only references work that exists
What people use it for
- A solo game developer who wants a first-draft sprint plan built from stories already marked Ready.
- A producer who needs a capacity-aware plan with a 20% buffer and a Must/Should/Nice split.
- A lead carrying unfinished stories into the next sprint who wants a clear carryover table with reasons and new estimates.
- A team checking a sprint's risks and Definition of Done before a sprint review.
Frequently asked questions
Is the sprint-plan skill free to use?
The source for this skill is released under the MIT licence, which is stated on this page. Check the licence file in the repository for the full terms before you reuse or redistribute it.
What do I need to run the sprint-plan skill?
You need Claude Code or a compatible SKILL.md agent, the Claude Code Game Studios repository structure with production/epics, design/gdd and hooks/yaml-helper.sh, and story files with a '> **Status**' line. Without a story backlog the skill stops and asks you to run /create-stories first.
Why does sprint-plan say no stories were found?
It only looks for files matching production/epics/**/story-*.md that contain a '> **Status**' line marked Ready. Check that your stories are in that location, use the story-NNN-slug.md naming, and have the Status line in the expected form.
Where do I install the sprint-plan skill?
Unzip the downloaded ZIP into ~/.claude/skills/ to use it across projects, or into a project's .claude/skills/ folder to use it in one project. In the Claude apps, upload it under Settings → Capabilities → Skills.
Does the skill need internet access?
No. The skill is listed as not using the network. It does run a shell script for config lookup and a directory listing, so review the SKILL.md file if you need to confirm what it runs.
How is this different from writing a sprint plan by hand?
The skill drafts the plan from your existing Ready stories, previous sprint and design docs, and fills in capacity, buffer, carryover and risks. It is not a substitute for a planning conversation, and it is not built for projects outside the Claude Code Game Studios structure.
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 sprint-plan if you run game sprints in the Claude Code Game Studios structure, already have stories marked Ready under production/epics and want a first-draft plan to edit. Pick another approach if your project has no story backlog, or if your team does not use that folder layout, since the skill expects specific folders, hooks and companion commands that this file does not provide.
Download the skill