Guide · Agent skill
How to use the Project Onboarding Guide agent skill (onboard)
Updated · written with AI from the project's own documentation and checked against our catalogue data
The onboard skill is an agent skill for Claude Code that writes a structured onboarding document for a new contributor or agent joining a game project. It is aimed at a new team member, or a team lead who wants the same orientation generated for each new person or agent. After this guide you will know what a skill is, how to install onboard, how to ask for it with a plain request, and what it will and will not do.
Project Onboarding GuideDonchitos/Claude-Code-Game-Studios · MIT Download the skill Quick start
- 1Get the ZIP. Download the onboard skill ZIP from this page. It contains SKILL.md.
- 2Unzip into a skills folder. Unzip into ~/.claude/skills/ for all projects, or into the project's .claude/skills/ for one project.
- 3Confirm the framework files. Check that CLAUDE.md, .claude/agents/, .claude/docs/ and hooks/yaml-helper.sh exist in the project.
- 4Ask for a role. Request an orientation with a role such as programmer, designer, narrative, QA, or production.
- 5Review the document. Read the generated onboarding document and check any NOT ASSESSED sections.
- 6Approve the save. Say yes only if you want the file written to production/onboarding/onboard-[role]-[date].md.
What an agent skill is and what onboard makes the agent do
An agent skill is a folder containing a SKILL.md file with instructions that an agent such as Claude Code loads when a request matches it. The onboard skill's description is an onboarding doc covering project state, conventions, and priorities for a specified role.
Given a role or area such as programmer, designer, narrative, QA, or production, the agent reads the project's CLAUDE.md, the matching agent definition in .claude/agents/, and the relevant project folders. It then writes a document covering the project summary, the role's responsibilities, key directories and files, conventions, current state, and first tasks.
- Project summary and your role
- Key directories and key files, with a read priority for each
- Conventions from CLAUDE.md and the agent definition
- Current state of your area, current sprint context, and common pitfalls
- First tasks and questions to ask
Install the onboard skill from the ZIP
Download the ZIP for the onboard skill from this page. The skill ships as a single file, SKILL.md. You can install it in one of two places.
For a personal install that applies to all your projects, unzip it into ~/.claude/skills/. For a single project, unzip it into that project's .claude/skills/ folder. In the Claude apps, you can instead upload the ZIP under Settings → Capabilities → Skills.
- Personal: ~/.claude/skills/
- Single project: the project's .claude/skills/ folder
- Claude apps: Settings → Capabilities → Skills, then upload the ZIP
Check the project layout and requirements before you run it
The skill depends on the Claude Code Game Studios framework files. It needs CLAUDE.md, .claude/agents/, .claude/docs/, and hooks/yaml-helper.sh. It needs Claude Code or a compatible agent runtime with access to the project's files. Git is optional and only adds recent-change context.
If CLAUDE.md or .claude/agents/ is missing, the skill reports that these files are missing from the install. It does not create them. Designer, narrative, QA, and production scans read the design/, tests/, and production/ folders, so those folders need to exist for useful output.
- Required: CLAUDE.md, .claude/agents/, .claude/docs/, hooks/yaml-helper.sh
- Required: Claude Code or compatible runtime with file access
- Optional: git, for recent commit history
Trigger the skill with a natural request
The skill takes an optional argument written as [role|area]. You can name the role in your request, for example programmer, designer, narrative, QA, or production. If you give no role, the skill produces a general orientation and suggests that a role would focus the next pass.
Phrase the request the way you would ask a colleague. The skill name is onboard, so mentioning onboarding for a role is enough to match it in most setups.
- Use /onboard programmer for a programmer orientation
- Use /onboard designer to focus on existing design documents
- Use /onboard with no argument for a general orientation
/onboard programmerUnderstand the output and where it is saved
The skill presents the onboarding document in the conversation. It then asks permission before writing anything. If you agree, it saves the file to production/onboarding/onboard-[role]-[date].md, creating the directory if needed.
If required inputs are missing, the skill stops and reports that it could not run, rather than guessing. For each input it records FOUND or ABSENT, and any section without data reads NOT ASSESSED — NO DATA.
- Asks before writing the file
- Writes to production/onboarding/ only after you say yes
- Reports NOT ASSESSED instead of filling gaps with guesses
Safety notes: scripts and network
The skill runs a script. Its allowed tools include a Bash call to hooks/yaml-helper.sh with the resolve_config argument, which is used to read the automation mode setting. It also has Read, Glob, Grep, and Write access. According to the assessment, it does not use the network.
Because it can write files, review the save prompt before agreeing. Because it runs a script from your project, check that hooks/yaml-helper.sh comes from the framework you trust.
- Runs scripts: yes, hooks/yaml-helper.sh via Bash
- Uses network: no
- Writes files only after asking permission
Tips for getting a useful onboarding document
The document is generated from project files and recent git history. It can be out of date or incomplete if those are stale, so check dates and open questions before you rely on it. Run it again after major changes instead of reusing an old file.
Give a specific role. A focused role produces a more useful document than a general orientation. Keep the saved files in production/onboarding/ so a team lead can reuse the same pattern for each new person.
- Refresh the document after big changes to design or code
- Choose the most specific role that matches the new contributor
- Read the open questions it lists before the first session
What people use it for
- Orienting a new programmer to a game's code root, key files, and conventions in CLAUDE.md
- Giving a new designer a map of existing design documents under design/
- Preparing a narrative writer by pointing them to world-building docs in design/narrative/
- Creating a repeatable onboarding file for each new agent or person on a team
- Getting a production lead's first-day summary of the current sprint and milestone under production/
Frequently asked questions
Is the onboard skill free to use?
The source material does not state a price or licence terms beyond the listing, which shows the licence as MIT. Check the repository's licence file to confirm the terms before reuse.
What do I need to install the onboard skill?
Download the ZIP from this page and unzip it into ~/.claude/skills/ or a project's .claude/skills/ folder. You can also upload it in the Claude apps under Settings → Capabilities → Skills. You also need Claude Code or a compatible agent runtime with access to the project files.
Does onboard work on any project?
No. It needs the Claude Code Game Studios layout, including CLAUDE.md, .claude/agents/, design/, and production/. On other projects the output will mostly report missing pieces.
Does the onboard skill use the internet?
According to the assessment, it does not use the network. It does run a script, hooks/yaml-helper.sh, through Bash to read the automation mode setting.
Why did onboard say it could not run?
A required input was missing, such as CLAUDE.md or the agent definition for the role. The skill reports NOT ASSESSED rather than guessing. Add the missing files from the framework and run it again.
How is onboard different from writing an onboarding doc by hand?
Onboard generates the document from project files and recent git history each time, using a fixed structure. A hand-written doc can be more current on decisions the files do not record, so it works well alongside the generated one.
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 onboard if you work in the Claude Code Game Studios framework and need a repeatable, role-specific orientation document generated from the project's own files. Pick something else if your project does not follow that folder layout, or if you need verified current facts, because the output is only as current as the files and git history it reads.
Download the skill