Guide · Agent skill
How to use the Project Stage Detection agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
Project Stage Detection is an agent skill for game development projects. You ask the agent where your project is in development, and it scans your folders for design docs, source code, prototypes, architecture notes and tests, then reports the likely stage, the gaps it can see, and suggested next steps. This guide covers what the skill does, how to install it, how to trigger it, what it needs, and where it can mislead you.
Project Stage DetectionDonchitos/Claude-Code-Game-Studios · MIT Download the skill Quick start
- 1Get the skill ZIP. Download the project-stage-detect ZIP from the page you are reading. The download contains SKILL.md.
- 2Unzip it into a skills folder. Unzip into ~/.claude/skills/ for personal use, or into your project's .claude/skills/ for one project.
- 3Add the required scripts. Make sure .claude/scripts/artifact-check.sh and .claude/hooks/yaml-helper.sh from the parent Claude-Code-Game-Studios repository are in the project, and that bash is available.
- 4Open the project in your agent. Start Claude Code, or your compatible agent, from the project root so relative paths resolve.
- 5Ask where the project stands. Type a plain question such as 'Where are we in development?' You can add a role like 'designer' to narrow the report.
- 6Read the stage and gaps. Check the observed stage, the evidence, and any disagreement with the configured stage. Treat a missing-code 'not assessed' message as an unknown, not as zero.
- 7Act on the next steps. Pick the gaps that matter for your current milestone and ask the agent to help with them, or re-run the skill after your changes.
What is an agent skill and what does this one do?
An agent skill is a folder of instructions and optional scripts that an AI coding agent such as Claude Code can load and follow when a request matches the skill's purpose. This one is defined in a file called SKILL.md.
When you ask where the project is in development, the skill runs a bundled checker script, artifact-check.sh, to see which expected artifacts exist. It then estimates the current stage, from concept through production, checks whether that inferred stage matches any configured stage, and lists gaps and next steps. It is built for a game studio workflow and reads files only, aside from possibly writing a report.
- Estimates the current stage (Concept, Systems Design, Technical Setup, Pre-Production, Production, and the explicit Polish and Release stages)
- Checks which expected artifacts are present, short or missing
- Compares the observed stage with the configured stage and flags disagreements
- Lists gaps and suggests next steps
How do you install the project-stage-detect skill?
There are two routes. Download the skill ZIP from the page you are reading, then either unzip it into your user skills folder or into a single project's skills folder. You can also upload it through the Claude apps.
For a personal install across projects, unzip into ~/.claude/skills/. For one project only, unzip into that project's .claude/skills/ folder. In the Claude apps, open Settings, go to Capabilities, and upload the skill under Skills.
- Personal install: unzip into ~/.claude/skills/
- Project install: unzip into your project's .claude/skills/
- Claude apps: Settings → Capabilities → Skills, then upload
Install the scripts the skill depends on
The skill does not work on its own. It calls two scripts from the parent Claude-Code-Game-Studios repository: .claude/scripts/artifact-check.sh and .claude/hooks/yaml-helper.sh. Without them, the artifact scan and config lookup cannot run.
Make sure these files are present in the project you are analysing, at the paths above, and that bash is available on your system. The skill's allowed commands match the relative path form, so keep the scripts where the skill expects them.
- .claude/scripts/artifact-check.sh from the parent repository
- .claude/hooks/yaml-helper.sh from the parent repository
- bash available on the system
How do you trigger the skill with a natural request?
You do not need special syntax. Ask the question in plain language. The skill is tuned to requests like 'Where are we in development?' and similar questions about project state, onboarding, or readiness for a milestone.
The skill accepts an optional role filter, such as 'programmer' or 'designer', so you can narrow the report to the work relevant to that role. Mention the role in your request if you want it.
- 'Where are we in development?'
- 'What's missing before our next milestone?'
- 'I just joined this project. What exists so far?'
- 'Check what design work we've skipped, from a designer's point of view.'
What does a good result look like?
Expect a report that names the observed stage, the evidence behind it, and any gaps. The skill runs its script first, so the existence checks come from a deterministic pass, and then adds judgement about content quality and what the script does not track.
The report can include a configured stage. If the configured stage and the observed artifacts disagree, the skill says so and lowers its confidence instead of simply repeating the configured value. Read that disagreement as a prompt to check, not as a verdict.
- Observed stage with the artifacts that support it
- Gaps, shown as missing or short items rather than a bare list of files
- Explicit note when the configured stage and observed stage disagree
- Suggested next steps
What are the safety notes?
The skill runs bash scripts, so it is not a purely read-only tool. It is designed to read files, and it can write a report if asked. It does not use the network. Review the allowed commands in SKILL.md before installing if your environment restricts shell access.
The skill's allowed tools are Read, Glob, Grep and Write, plus the bash calls to artifact-check.sh and yaml-helper.sh. Any other shell command is outside its permission grant.
- Runs scripts: yes (bash only, the two bundled scripts)
- Uses network: no
- Writes files: possibly, for a report
- Reads project files: yes
Tips for getting accurate results
The skill only sees what is in the folder it analyses. If design work, prototypes or code live in another tool, a wiki or a different repository, the report can show an earlier stage than reality. Keep the repository current or treat the stage as a lower bound.
Run it from the project root so the relative script paths resolve. If the source code root cannot be found, the skill reports that the code was not assessed rather than treating the project as empty, so check that message before drawing conclusions.
- Run from the project root
- Keep design docs and architecture notes in the repository
- Treat a 'not assessed' message as missing information, not as zero files
- Re-run after adding documents to see the gap list update
What people use it for
- Onboarding onto an existing game project to get a quick map of what exists
- Checking for missing design or architecture documents before a milestone or stage gate
- Solo developers checking whether they skipped design or architecture work
- Leads reviewing project state across design, code, prototypes and tests
Frequently asked questions
How do I install the project-stage-detect skill?
Download the skill ZIP from the page, then unzip it into ~/.claude/skills/ for personal use or into the project's .claude/skills/ folder for one project. You can also upload it in the Claude apps under Settings → Capabilities → Skills.
What does the project-stage-detect skill need to run?
It needs Claude Code or a compatible agent with file-read and shell permissions, the bundled artifact-check.sh and yaml-helper.sh scripts from the parent Claude-Code-Game-Studios repository, and bash on the system. It does not use the network.
Is the project-stage-detect skill free?
The skill is licensed under the MIT licence. The source does not list prices for it, so check the parent repository and your agent's own terms for anything beyond the licence.
Can I use it on a non-game project?
It is not suited to that. The stage model and artifact list are built around game development folders, so results for other kinds of projects will be misleading.
Why does the skill say my project is at an earlier stage than I expect?
It only sees what is inside the project folder. Work kept in other tools or repositories is invisible to it, so it may report an earlier stage than reality. Check whether your design or code lives outside the repo.
Does the skill change my project files?
It is built to read files only. It may write a report if you ask for one, and its allowed tools include Write for that purpose.
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 lead who wants a fast, file-based map of what a game project contains and what is missing before a milestone. Pick something else if your project is not a game, or if most of your work lives outside the repository, because the skill will only report what it can see.
Download the skill