How to use the Bootstrap Soul agent skill

What is an agent skill and what does Bootstrap Soul do?

An agent skill is a folder of instructions and supporting files that an AI agent reads when it needs to handle a particular kind of task. The agent loads the skill only when a request matches it.

Bootstrap Soul is the onboarding skill for creating an AI partner's identity. It runs a conversation of 5 to 8 rounds across four phases: Hello (preferred language), You (your role, pain points, the assistant's name and relationship framing), Personality (traits, communication style, autonomy and pushback preference), and Depth (aspirations, failure philosophy, boundaries).

• Asks one to three questions per round rather than a long form
• Drafts a SOUL.md under 300 words in English, regardless of the conversation language
• Shows the draft for your confirmation before saving anything
• Saves the file only through the setup_agent tool

How do you install the Bootstrap Soul skill?

Download the bootstrap ZIP file from this page and unzip it. The skill folder contains SKILL.md, a templates folder with SOUL.template.md, and a references folder with conversation-guide.md.

You can install it for yourself or for a single project. For a personal install, place the unzipped bootstrap folder inside ~/.claude/skills/. For a project install, place it inside the project's .claude/skills/ folder.

In the Claude apps, you can instead upload the skill under Settings → Capabilities → Skills.

• Personal: ~/.claude/skills/bootstrap/
• Project: .claude/skills/bootstrap/ inside your project
• Claude apps: Settings → Capabilities → Skills

How do you trigger the skill with natural requests?

You do not need a special command. Describe what you want in plain language and the agent should recognise the request. The skill's description lists trigger phrases such as "create my SOUL.md", "bootstrap my agent", "set up my AI partner", "define who you are", "let's do onboarding", and "personalize this AI".

The skill also triggers for updates, such as "update my SOUL.md", "change my AI's personality", or "tweak the soul". It can also trigger when a SOUL.md is missing.

• "Create my SOUL.md"
• "Set up my AI partner"
• "Let's do onboarding"
• "Change my AI's personality"

What happens during the conversation?

The agent opens with a brief multilingual greeting and asks which language to use. It then asks about your role, what you want handled for you, and what to call the AI. Later rounds ask how the assistant should behave, how it should handle pushback and mistakes, and what boundaries you want.

The skill is written to react conversationally rather than read questions off a list. If you answer briefly, the agent asks follow-up questions. If you answer at length, it moves on faster. Once the draft is ready, the agent presents it and asks whether it feels right. It keeps revising until you confirm.

What does the skill need to run?

The skill depends on a setup_agent tool being available to the agent runtime. SKILL.md calls this tool to save the final file, but the tool is not defined in the provided files. The skill also forbids writing SOUL.md by hand with shell tools, so without setup_agent it cannot complete its final step.

According to the source, the skill runs no scripts and makes no network calls.

• A setup_agent tool available in your agent runtime
• No scripts are run by the skill
• No network use is stated in the skill

What are the safety and privacy notes?

The skill stores your personal details, role, and preferences in SOUL.md through the setup tool. If you do not want that information kept in a persistent file, do not use this skill.

If setup_agent returns an error, the skill tells the agent to report the error to you and not claim success. Check that the saved SOUL.md contains what you confirmed.

• Your name, role and preferences are written into a persistent SOUL.md file
• The skill does not allow manual file writes as a workaround

What tips help you get a better SOUL.md?

Answer with concrete details. The skill turns your pain points into what the assistant should do, and your word choices into how it talks. Vague answers produce generic output.

Review the draft carefully before confirming. Core traits are written as behavioural rules, so check that each one reflects how you want the assistant to act. To change the persona later, ask directly, for example "update my SOUL.md".

• Give specific examples of tasks you want handled
• Match the tone you want the assistant to use
• Check each core trait before confirming