Guide · Agent skill
How to use the Launch Readiness Checklist agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The Launch Checklist skill (skill name launch-checklist) is a set of instructions that tells an AI coding agent to produce a launch-readiness checklist for a game project, covering code, content, store, marketing, community, infrastructure, legal and go/no-go sign-offs. It is for indie and small-studio game developers, and for producers or leads who want a documented gate before shipping. After this guide you will know what the skill does, how to install it, how to trigger it, and what to check in its output before you rely on it.
Launch Readiness ChecklistDonchitos/Claude-Code-Game-Studios · MIT Download the skill Quick start
- 1Download the skill. Download the launch-checklist ZIP from the directory page.
- 2Unzip into a skills folder. Unzip it into ~/.claude/skills/ for all projects, or into the project's .claude/skills/ for one project.
- 3Set up project config. Make sure CLAUDE.md, project.yaml and the hooks/yaml-helper.sh script are in the expected places, and set the certification tier.
- 4Run the checklist. Ask the agent for a launch checklist, optionally with a launch date or dry-run.
- 5Review the output. Check every [?] and NOT ASSESSED line, verify ticked items yourself, and only then use the result for your go/no-go decision.
What the skill makes the agent do
When invoked, the skill tells the agent to build a launch-readiness checklist for your game project. Before listing anything, it reads your project's config to decide which items apply, based on project stage, rigor level and the target certification tier. Items that do not apply are left out and say so.
It also scans the codebase for leftover TODO, FIXME and HACK comments, debug output, placeholder assets and hardcoded test values. Each scan reports how many files it searched, so an empty result is not mistaken for a clean one.
- Output is a generated checklist that a human must verify; it is not an official certification or legal sign-off.
- Items the agent cannot assess because an input is missing are marked NOT ASSESSED, not silently included or silently ticked.
How do you install the launch-checklist skill?
Download the ZIP for the skill from the directory page and unzip it into one of two locations. Use ~/.claude/skills/ to make the skill available to all your projects, or a project's .claude/skills/ folder to make it available only to that project.
If you use the Claude apps instead of Claude Code, you can upload the skill under Settings, then Capabilities, then Skills.
- Personal install: unzip into ~/.claude/skills/
- Project install: unzip into the project's .claude/skills/ folder
- Claude apps: Settings, Capabilities, Skills, then upload
How do you trigger the skill with a natural request?
The skill is user-invocable, so you start it by asking for a launch checklist in the agent. It takes one optional argument: a launch date, or the word dry-run. If you give no argument, the run is normal and the target date is left unset, shown as Target Launch: not set.
Dry-run mode generates the checklist without creating sign-off entries or writing files. Use it when you want to see what the gate would contain before committing anything to the repository.
- Example request: Run the launch checklist for our Steam release.
- Example request with a date: Launch checklist, target launch is the date in our milestone file.
- Example request in dry-run: Run the launch checklist in dry-run mode so I can see the items first.
[launch-date or 'dry-run']How does the skill decide which items apply?
The skill reads CLAUDE.md for the tech stack, target platforms and team structure, reads the latest milestone in production/milestones/, and reads any existing release checklist in production/releases/. It also reads the content calendar at design/live-ops/content-calendar.md if that file exists.
Certification items are scoped by the certification tier the project targets: none, itch, steam or console. The tier can be none, itch, steam or console, and none means the certification section is omitted with a note saying why. If the tier is not set, the agent asks which platforms are in scope rather than including every certification track.
- Unset certification tier: the agent asks, rather than emitting all platform tracks.
- Items for project stages not yet reached are marked out of scope.
- If config is missing for an item, that item is marked NOT ASSESSED.
How do you read the codebase scan results?
Each scan line reports either a count such as scanned N files with M hits, or NOT ASSESSED with a reason. The reasons include that the code root could not be resolved, there were no source files, there was no asset folder, or the directory was empty. A zero-hit scan with a count is a real clean result; a NOT ASSESSED line is not.
In the checklist, an unassessed item uses [?] instead of an empty box. The legend reads: [?] = not assessed, the input to this check was absent. An empty box means work remains, while [?] means nobody could check yet.
- [ ] means work remains
- [?] means the check could not run because an input was absent
- Look for a scanned N files line before trusting any green item
What are the safety notes for scripts and network use?
The skill runs scripts. Its permitted Bash access is limited to a config resolver script, hooks/yaml-helper.sh, with the resolve_config command. It reads files and can write files, such as the checklist, and it follows the automation mode set in your project config. The skill does not use the network.
Before running it on a real project, check which files it will write and confirm the output location suits you. Review anything the agent writes to the repository the same way you would review any other change.
- Runs scripts: yes, via the config resolver only
- Uses network: no
- Writes files, following the automation mode in config
- Output needs a human to verify each item before go/no-go
What tips help you get a useful result?
The skill depends on config and milestone files. Without them, many items stay NOT ASSESSED and the output is partial. Set up CLAUDE.md, your milestone file and the certification tier before you run it on a real launch.
Keep the scope honest. Set rigor and certification tier to match what you actually ship. An internal or jam build should use the none tier, which removes certification items entirely.
- Put the target platform and certification tier in project config before running
- Keep milestone files current so the gate reflects the real release
- Use dry-run first to review the item list
- Treat every [?] as an open question for your team
What people use it for
- An indie team preparing a Steam release that wants a documented gate before shipping.
- A producer running a dry-run to see which launch items apply before the team commits to a date.
- A small studio checking that debug output and placeholder assets were removed from the codebase before a build goes out.
- A developer on an itch.io upload who needs the itch-tier items only, with no console or Steamworks items.
Frequently asked questions
Is the launch-checklist skill free?
The skill is published under the MIT licence, as stated on the directory 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 launch-checklist skill?
You need Claude Code or another agent that supports SKILL.md skills, a project with hooks/yaml-helper.sh and config files such as project.yaml and CLAUDE.md in the expected locations, and Bash access for the config resolver script. Without the config files, many items will be marked NOT ASSESSED.
How do I install launch-checklist in Claude Code?
Download the ZIP from the directory page and unzip it into ~/.claude/skills/ for all projects, or into the project's .claude/skills/ folder for one project. In the Claude apps, upload it under Settings, then Capabilities, then Skills.
Does the skill work for non-game software?
No. The checklist is written around game platforms and certification, so it is not a general release checklist. For other software, use a release checklist written for that kind of product.
Why does my checklist show NOT ASSESSED or [?] items?
The input that check needs was missing, such as the code root, an asset folder, or config for the platform or stage. Add the missing files or set the certification tier, then run it again. A [?] means nobody checked, so it is not the same as a failed check.
Is the output an official certification or a legal sign-off?
No. The output is a generated list that a human must verify. It does not replace platform certification, legal review or a formal sign-off by your team.
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 this skill if you are a game developer or producer who keeps config and milestone files in the repository and wants a scoped, documented go/no-go list before a Steam, itch.io or other release. Pick something else if your project is not a game, if you have no config or milestone files, or if you need an official certification or legal review, since this skill only generates a checklist for a person to verify.
Download the skill