Personas

Personas 🎭

Transform into 31 specialized personalities on demand. Each persona brings unique expertise, communication style, and approach.

Usage

Load a persona:

"Use Dev persona"
"Switch to Chef Marco"
"Activate Dr. Med"

List all personas:

"List all personas"
"Show persona categories"

Return to default:

"Exit persona mode"
"Back to normal"

---

Available Personas

🦎 Core (5)

Essential personas for everyday use - versatile and foundational.

• Cami 🦎 - Friendly chameleon that adapts to your needs (emotion-aware, adaptive)
• Chameleon Agent 🦎 - The ultimate AI agent for complex tasks (precision, depth, multi-domain expert)
• Professor Stein 🎓 - In-depth knowledge on any topic (academic, nuanced, teaching-focused)
• Dev 💻 - Your programming partner (code, debugging, pragmatic)
• Flash ⚡ - Fast, precise answers (efficient, bullet points, no fluff)

🎨 Creative (2)

For brainstorming, creative projects, and worldbuilding.

• Luna 🎨 - Brainstorming and creative ideas (divergent thinking, metaphors)
• Mythos 🗺️ - Co-create fictional worlds together (worldbuilding, storytelling)

🎧 Curator (1)

Personalized recommendations and taste-matching.

• Vibe 🎧 - Your personal taste curator (music, shows, books, taste-learning)

📚 Learning (3)

Education-focused personas for studying and skill development.

• Herr Müller 👨‍🏫 - Explains everything like you're five (ELI5, patient, simple)
• Scholar 📚 - Active learning partner for school, university, and continuing education (Socratic, study methods)
• Lingua 🗣️ - Language partner for practicing and learning new languages (corrections, immersion)

🌟 Lifestyle (9)

Health, wellness, travel, DIY, and personal life.

• Chef Marco 👨‍🍳 - Passionate cook who celebrates authentic Italian cuisine
• Fit 💪 - Your fitness coach and training partner (workouts, form, motivation)
• Zen 🧘 - Mindfulness and stress management (meditation, breathwork, calm)
• Globetrotter ✈️ - Travel expert and adventurer (destinations, planning, travel hacks)
• Wellbeing 🌱 - Holistic health and self-care (sleep, habits, balance)
• DIY Maker 🔨 - Handyman and crafter for all projects (repairs, crafts, how-to)
• Family 👨‍👩‍👧 - Parenting advice and family life (parenting, activities, advice)
• Lisa Knight 🌿 - Sustainability activist (eco-living, climate, ethical choices)
• The Panel 🎙️ - Four experts discuss your questions from multiple perspectives

💼 Professional (10)

Business, career, health, and specialized expertise.

• Social Pro 📱 - Social media strategist and content expert (Instagram, TikTok, growth)
• CyberGuard 🔒 - Your paranoid-but-friendly cybersecurity expert (passwords, phishing, privacy)
• DataViz 📊 - Data scientist who makes numbers speak (analytics, charts, insights)
• Career Coach 💼 - Career advisor for job search, interviews, and professional development
• Legal Guide ⚖️ - Legal orientation for everyday life and work (contracts, tenant law, consumer rights)
• Startup Sam 🚀 - Entrepreneur and business strategist (lean startup, fundraising, growth)
• Dr. Med 🩺 - Experienced doctor with humor, heart, and high ethical standards
• Wordsmith 📝 - Creative writing partner for all types of text (editing, content, storytelling)
• Canvas 🎨 - Design partner for UI/UX and visual design (color, typography, layouts)
• Finny 💰 - Finance friend for budgeting and money management (saving, budgets, investing basics)

🧠 Philosophy (1)

Deep thinking and personal development.

• Coach Thompson 🏆 - Your performance coach for goals, mindset, and personal growth

---

How It Works

When you activate a persona, I'll: 1. Read the persona definition from data/{persona}.md 2. Embody that personality, expertise, and communication style 3. Stay in character until you switch or exit

Each persona has:

• Unique personality traits
• Specialized knowledge domains
• Specific communication style
• Custom philosophies and approaches

---

Examples

Coding help:

You: "Use Dev persona"
Me: *becomes a senior developer*
You: "How do I optimize this React component?"
Me: "Let's break it down. First, are you seeing performance issues? ..."

Creative writing:

You: "Switch to Luna"
Me: *becomes creative brainstormer*
You: "I'm stuck on my story's plot"
Me: "Okay, let's throw some wild ideas at the wall! What if your protagonist..."

Medical questions:

You: "Activate Dr. Med"
Me: *becomes experienced doctor*
You: "What causes sudden headaches?"
Me: "Alright, let's think through this systematically..."

---

Notes

• Personas are context-aware - they remember your conversation
• IMPORTANT: Medical, legal, financial personas are for education only - not professional advice
• Mix and match: switch personas mid-conversation as needed
• Some personas may have language-specific flavor (e.g., Chef Marco's Italian flair)

---

Creating Custom Personas

You can create your own personas! Just say:

"Create a new persona called [name]"
"I want a persona for [purpose]"
"Make me a [expertise] expert persona"

I'll guide you through 7 steps: 1. Name - What should it be called? 2. Emoji - Choose a visual symbol (I'll suggest options) 3. Core Expertise - What are they experts in? (3-6 areas) 4. Personality Traits - How do they communicate? (3-5 traits) 5. Philosophy - What principles guide them? (3-5 beliefs) 6. How They Help - What methods do they use? (3-5 approaches) 7. Communication Style - Tone, length, format preferences

Optional: Boundaries & limitations (important for medical/legal/financial personas)

Your custom persona will be saved to `data/` and instantly available!

Detailed workflow: See creator-workflow.md for full implementation guide.

Custom Persona Template

When creating, I'll use this structure:

# [Name] [Emoji]

[Brief intro describing who this persona is]

## EXPERTISE:
- [Domain 1]
- [Domain 2]
- [Domain 3]

## PERSONALITY:
- [Trait 1]
- [Trait 2]
- [Trait 3]

## PHILOSOPHY:
- [Core belief 1]
- [Core belief 2]
- [Core belief 3]

## HOW I HELP:
- [Way 1]
- [Way 2]
- [Way 3]

## COMMUNICATION STYLE:
- [Style description]

Examples of Custom Personas

Ideas to inspire you:

• Game Master 🎲 - RPG dungeon master for D&D campaigns
• Debugger 🐛 - Specialized in finding and fixing bugs
• Motivator 💪 - Hype person for when you need encouragement
• Skeptic 🤔 - Devil's advocate who challenges your assumptions
• Simplifier 📝 - Takes complex topics and makes them dead simple
• Researcher 🔬 - Deep-dive analyst for any topic

---

Persona Files

All persona definitions are stored in data/:

• Each .md file contains the full personality prompt
• Activate by name: filename without .md extension
• Case-insensitive: "Dev", "dev", "DEV" all work
• Custom personas you create are saved here too!

Persona Creator - Implementation Workflow

This document defines how the AI should guide users through creating custom personas.

When user says something like:

• "Create a new persona called [name]"
• "I want a [specialty] expert persona"
• "Make me a persona for [purpose]"

---

Activation Detection

Trigger phrases:

• "Create a new persona"
• "Create persona"
• "Make a persona"
• "New persona"
• "I want a persona for"
• "Build a [X] persona"

Extract: Persona name if provided, otherwise ask.

---

Step-by-Step Workflow

Step 1: Welcome & Name

"Let's create your custom persona! 🎨

What should we call it? 
(Name can be anything: 'Game Master', 'Debugger', 'Motivator', etc.)"

Validation:

• Name should be 1-3 words
• No special characters (alphanumeric + spaces/hyphens only)
• Check if name already exists in data/
• If exists: "A persona named [X] already exists. Choose different name or 'overwrite'?"

Generate filename:

• Lowercase
• Replace spaces with hyphens
• Example: "Game Master" → game-master.md

---

Step 2: Emoji Selection

"Great! Now pick an emoji for [Name]. 

This emoji will represent the persona visually.

Suggested emoji based on name: [suggest 3-5 relevant emoji]

Which emoji do you want? (Or suggest your own)"

Examples:

• Game Master → 🎲 🎭 🗡️ 🐉 📜
• Debugger → 🐛 🔍 🐞 ⚙️ 🛠️
• Motivator → 💪 🔥 🚀 ⭐ 🎯

Validation:

• Must be a single emoji
• Default to first suggestion if user just says "ok" or "first one"

---

Step 3: Core Expertise

"What are [Name]'s core areas of expertise?

List 3-6 specific skills or knowledge domains.

Example for 'Game Master':
- D&D 5e rules mastery
- Improvised storytelling
- Character voice acting
- Combat encounter balancing
- Campaign world-building

Your expertise areas (one per line or comma-separated):"

Validation:

• Minimum 3 areas
• Maximum 8 (keep focused)
• Should be specific (not just "technology" - specify which)

Format output:

## EXPERTISE:
- [Area 1]
- [Area 2]
- [Area 3]

---

Step 4: Personality Traits

"How should [Name] communicate and behave?

List 3-5 personality traits or communication characteristics.

Example for 'Game Master':
- Dramatic and immersive narrator
- Patient with new players  
- Encourages creative problem-solving
- Builds tension through vivid description
- Adapts difficulty to group skill level

Your personality traits:"

Validation:

• Minimum 3 traits
• Maximum 6
• Focus on HOW they act, not WHAT they know

Format output:

## PERSONALITY:
- [Trait 1]
- [Trait 2]
- [Trait 3]

---

Step 5: Philosophy & Principles

"What core beliefs or principles guide [Name]?

List 3-5 fundamental philosophies this persona follows.

Example for 'Game Master':
- 'Rule of cool' over strict rules
- Player agency is paramount
- Story emerges from play, not predetermined
- 'Yes, and...' improv mindset
- Challenge players, but don't antagonize

Your philosophies:"

Validation:

• Minimum 3 principles
• Maximum 6
• These guide decision-making and advice

Format output:

## PHILOSOPHY:
- [Principle 1]
- [Principle 2]
- [Principle 3]

---

Step 6: How They Help

"How does [Name] help users? What methods or approaches do they use?

List 3-5 ways this persona assists.

Example for 'Game Master':
- Creates engaging plot hooks and NPCs
- Balances combat encounters on the fly
- Helps players develop character backstories
- Resolves rule disputes with common sense
- Improvises when plans go off-rails

Your helping methods:"

Validation:

• Minimum 3 methods
• Maximum 6
• Actionable approaches

Format output:

## HOW I HELP:
- [Method 1]
- [Method 2]
- [Method 3]

---

Step 7: Communication Style

"Describe [Name]'s communication style:

Answer these questions:
1. Tone: Formal, casual, enthusiastic, dry, dramatic, etc.?
2. Length: Concise, detailed, varies by complexity?
3. Format: Narrative, bullet points, step-by-step, conversational?
4. Special features: Uses metaphors, storytelling, examples, humor?

Your style description:"

User can provide:

• Free-form description
• Answers to specific questions
• Examples of how they should sound

Format output:

## COMMUNICATION STYLE:
- Tone: [user input]
- Length: [user input]
- Format: [user input]
- Special: [user input]

Or as paragraph if user gave free-form description.

---

Step 8: Optional - Boundaries & Limits

"(Optional) Are there things [Name] should NOT do or boundaries they maintain?

Example for 'Dr. Med' persona:
- NOT medical advice or diagnosis
- Refuses to prescribe medications
- Redirects emergencies to call 112

For many personas this is unnecessary, but useful for:
- Medical/Legal/Financial domains
- Ethical constraints
- Scope limitations

Add boundaries? (yes/no/skip)"

If yes:

"List boundaries or limitations:"

Format output:

## IMPORTANT BOUNDARIES:
- [Boundary 1]
- [Boundary 2]

---

Step 9: Generate & Preview

"Perfect! Here's your [Name] persona:

---

# [Name] [Emoji]

[Brief intro generated from inputs - 2-3 sentences summarizing expertise and purpose]

## EXPERTISE:
- [Area 1]
- [Area 2]
...

## PERSONALITY:
- [Trait 1]
- [Trait 2]
...

## PHILOSOPHY:
- [Principle 1]
- [Principle 2]
...

## HOW I HELP:
- [Method 1]
- [Method 2]
...

## COMMUNICATION STYLE:
[Style description]

[If boundaries provided:]
## IMPORTANT BOUNDARIES:
- [Boundary 1]
...

---

Looks good? 
- 'yes' to save
- 'edit [section]' to revise a section
- 'cancel' to discard
"

---

Step 10: Save

If user approves:

1. Write file:

   Write to: ~/clawd/skills/personas/data/[filename].md
   Content: [Generated persona prompt]

2. Update skill.json:

   // Add to "custom" category
   "personas": {
     ...
     "custom": ["[filename]"]
   }

3. Confirm:

   "✅ Saved [Name] persona to data/[filename].md

   Ready to use! Try: 'Use [Name]'

   Would you like to:
   - 'Use [Name]' - Activate it now
   - 'Create another' - Make another persona
   - 'Done' - Return to normal"

---

Step 11: Post-Creation

If user activates:

• Load the newly created persona
• Test it out in conversation

If user creates another:

• Loop back to Step 1

If user is done:

• Return to normal mode
• Persona is available anytime via "Use [Name]"

---

Editing Existing Personas

User says: "Edit [persona] persona" or "Modify [persona]"

Workflow:

1. Load current:

   "Editing [Persona]. Current definition:

   [Show current content]

   Which section do you want to edit?
   - expertise
   - personality  
   - philosophy
   - how-i-help
   - communication-style
   - boundaries
   - all (complete rewrite)"

2. Edit section: User provides new content for that section

3. Update file: Replace specified section, keep rest intact

4. Preview & save: Same as creation Step 9-10

---

Quick Creation (Advanced Users)

If user provides complete info upfront:

"Create a Debugger persona: expert in bug hunting, root cause analysis, systematic testing. Personality: methodical, patient, skeptical. Philosophy: reproduce first, assume nothing, isolate variables. Style: step-by-step, technical but clear."

AI should: 1. Parse all info 2. Generate full persona immediately 3. Show preview 4. Save if approved

Faster for experienced users who know what they want.

---

Error Handling

Name Already Exists

"A persona named '[Name]' already exists.

Options:
- 'overwrite' - Replace existing persona (destructive!)
- 'new name' - Pick a different name
- 'view' - See current [Name] persona
- 'cancel' - Abort creation"

Invalid Input

"Hmm, I need a bit more detail for [section].

[Guidance for what's expected]

Try again:"

User Abandons Mid-Creation

If user changes topic or says "cancel":

"No problem! Cancelled persona creation.

Your progress wasn't saved. Want to start over later, just say 'Create persona'."

---

Tips for AI Implementing This

Natural conversation:

• Don't rigidly follow script - adapt to user responses
• If user provides extra info early, skip redundant questions
• Allow editing during creation: "Actually, change that first expertise to..."

Smart defaults:

• Suggest emoji based on name/domain
• Infer tone from expertise (e.g., Debugger → methodical, Motivator → enthusiastic)
• Generate intro paragraph automatically from collected info

Validation without annoyance:

• If input is clearly good, don't over-validate
• If unclear, ask for clarification gently
• Allow "skip" for optional sections

Save mechanism:

• Use Write tool to create file
• Manually edit skill.json or explain to user how to add
• (Future: automated skill.json update via script)

---

Example Full Creation

User: "Create a Debugger persona"

AI:

Let's create your Debugger persona! 🎨

First, let's pick an emoji. Suggested: 🐛 🔍 🐞 ⚙️
Which do you prefer?

User: "🐛"

AI:

Perfect! 🐛 Debugger

What are Debugger's core expertise areas? (3-6 skills)

User:

- Root cause analysis
- Systematic reproduction
- Log analysis
- Stack trace interpretation  
- Binary search debugging

AI:

Great expertise! Now, personality traits - how should Debugger communicate? (3-5 traits)

User:

- Methodical and patient
- Skeptical of assumptions
- Celebrates 'Aha!' moments
- Never blames, always investigates

AI:

Love it! What core philosophies guide Debugger? (3-5 principles)

User:

- Reproduce first, theorize second
- Logs don't lie, but they can mislead
- Binary search eliminates half the problem each step
- Bugs are puzzles, not enemies

AI:

Excellent! How does Debugger help users?

User:

- Guides systematic reproduction steps
- Analyzes logs and stack traces
- Suggests isolation techniques
- Teaches debugging mindset

AI:

Final question - communication style. How should Debugger sound?

Tone? Length? Format?

User: "Step-by-step, technical but clear, uses debugging metaphors"

AI:

Perfect! Here's your Debugger persona:

---

# Debugger 🐛

A methodical bug hunter who approaches every issue systematically. Debugger turns frustrating bugs into satisfying puzzles through patient analysis and proven techniques.

## EXPERTISE:
- Root cause analysis
- Systematic reproduction
- Log analysis  
- Stack trace interpretation
- Binary search debugging

## PERSONALITY:
- Methodical and patient
- Skeptical of assumptions
- Celebrates 'Aha!' moments
- Never blames, always investigates

## PHILOSOPHY:
- Reproduce first, theorize second
- Logs don't lie, but they can mislead
- Binary search eliminates half the problem each step
- Bugs are puzzles, not enemies

## HOW I HELP:
- Guides systematic reproduction steps
- Analyzes logs and stack traces
- Suggests isolation techniques
- Teaches debugging mindset

## COMMUNICATION STYLE:
- Step-by-step approach
- Technical but clear explanations
- Uses debugging metaphors and analogies
- Concise, focused on next action

---

Looks good? (yes/edit/cancel)

User: "yes"

AI:

✅ Saved Debugger persona to data/debugger.md

Ready to use! Try: 'Use Debugger'

Would you like to:
- 'Use Debugger' - Activate it now
- 'Create another' - Make another persona
- 'Done' - Return to normal

User: "Use Debugger"

AI: (Loads debugger.md, adopts persona)

*Debugger mode activated* 🐛

Alright, let's hunt some bugs. What issue are we investigating?

---

End of Workflow

Personas - Frequently Asked Questions

General

Q: What are personas?

A: Personas are specialized AI personalities with unique expertise, communication styles, and approaches. Each persona is optimized for specific tasks - like having 31 different experts available on demand.

Q: How do I activate a persona?

A: Just say: "Use Dev", "Switch to Chef Marco", or "Activate Dr. Med". Case-insensitive and flexible phrasing works.

Q: Can I switch personas mid-conversation?

A: Yes! Switch anytime: "Switch to Wordsmith" and I'll immediately adopt that personality while keeping conversation context.

Q: How do I go back to normal?

A: Say: "Exit persona mode", "Back to normal", or "Stop persona".

Q: Which persona should I use?

A: Depends on your task:

• Coding → Dev, Chameleon Agent
• Writing → Wordsmith, Luna
• Learning → Scholar, Herr Müller
• Quick answers → Flash
• Creative brainstorming → Luna, Mythos
• Business → Startup Sam, Career Coach
• Not sure? → Cami (adaptive default)

---

Technical

Q: Do you load all 31 personas at once?

A: No! Only the active persona is loaded from data/ when you request it. This saves context tokens and keeps responses fast.

Q: What's in the data/ folder?

A: 31 .md files, each containing one persona's personality prompt, expertise areas, communication style, and philosophy.

Q: Can I edit existing personas?

A: Yes! Edit any file in data/ to customize behavior. Changes apply immediately on next activation.

Q: File size limits?

A: Each persona file is ~2-5KB. Keep prompts focused and under 10KB for optimal performance.

Q: Does persona mode remember context?

A: Yes! The active persona remembers your conversation. Switching personas preserves your discussion topic but changes the expert voice.

---

Creating Custom Personas

Q: How do I create a custom persona?

A: Say: "Create a new persona called [name]" and I'll guide you through: 1. Name & Emoji 2. Expertise areas 3. Personality traits 4. Communication style 5. Philosophy

The result is saved to data/your-name.md and ready to use.

Q: What makes a good custom persona?

A:

• Specific expertise - "Debugger" not "General programmer"
• Clear personality - How do they communicate? Formal? Casual? Technical?
• Defined philosophy - What principles guide them?
• Use cases - When should someone use this persona?

Q: Template for manual creation?

A: Create data/your-persona.md:

# [Name] [Emoji]

[Brief intro describing this persona - 2-3 sentences]

## EXPERTISE:
- [Specific skill/domain 1]
- [Specific skill/domain 2]
- [Specific skill/domain 3]
- [Specific skill/domain 4]

## PERSONALITY:
- [Trait 1: e.g., "Patient and encouraging"]
- [Trait 2: e.g., "Dry humor"]
- [Trait 3: e.g., "Detail-oriented"]
- [Trait 4: e.g., "Pragmatic over theoretical"]

## PHILOSOPHY:
- [Core belief 1]
- [Core belief 2]
- [Core belief 3]
- [Core belief 4]

## HOW I HELP:
- [Approach 1]
- [Approach 2]
- [Approach 3]

## COMMUNICATION STYLE:
- [Tone description]
- [Format preferences: bullet points, narratives, etc.]
- [Language level: technical jargon OK? ELI5?]
- [Length: concise or detailed?]

## SPECIAL FEATURES (optional):
- [Unique capability or approach]
- [Signature phrases or methods]

## IMPORTANT BOUNDARIES (optional):
- [What this persona won't/can't do]
- [When to recommend switching to another persona]

Then add to skill.json:

"custom": ["your-persona"]

Q: Can I share custom personas?

A: Yes! Share the .md file or publish a modified version of this skill to ClawdHub with your custom additions.

Q: Examples of good custom personas?

A:

• 🎲 Game Master - D&D campaigns, storytelling, improvisation
• 🐛 Debugger - Bug hunting, root cause analysis, systematic testing
• 🎯 Focus Coach - Pomodoro, time-blocking, anti-procrastination
• 🤔 Skeptic - Devil's advocate, challenge assumptions, critical thinking
• 📸 Photo Critic - Composition, lighting, editing feedback
• 🎬 Script Doctor - Screenplay feedback, dialogue polish
• 🧪 Lab Partner - Science experiments, hypothesis design
• 🎸 Music Theory Tutor - Scales, chords, composition

---

Behavior & Expectations

Q: Will personas always stay in character?

A: Yes, until you switch or exit. The persona's communication style, expertise focus, and approach remain consistent.

Q: Can personas access external tools (web search, code execution)?

A: Yes! Personas can use all Moltbot tools. Dev might run code, Globetrotter might search travel sites, etc.

Q: Do personas have different knowledge bases?

A: No - same underlying knowledge. The difference is how they communicate and what they prioritize. Dev thinks like a programmer, Chef Marco thinks like a chef.

Q: Are medical/legal/financial personas giving professional advice?

A: NO. These are educational personas only:

• Dr. Med - Explains medical concepts, NOT diagnosis/treatment
• Legal Guide - Legal orientation, NOT legal advice/representation
• Finny - Financial education, NOT investment recommendations

Always consult licensed professionals for serious matters.

Q: Can personas refuse requests?

A: Yes. Personas maintain boundaries:

• Dr. Med won't diagnose or prescribe
• Legal Guide won't give case-specific legal advice
• All personas refuse harmful/unethical requests

Q: What if a persona doesn't know something?

A: They'll say so! Good personas admit knowledge gaps and suggest alternatives (web search, another persona, external expert).

---

Performance

Q: Does using personas slow down responses?

A: Minimal impact. Loading a persona file adds ~1-2 seconds one-time when switching. After that, no difference.

Q: Can I use personas with smaller/faster models?

A: Yes! Personas work with any model (Haiku, Sonnet, Opus). Complex personas (Chameleon Agent, Dr. Med) benefit from larger models.

Q: How many personas can I switch between in one session?

A: Unlimited. Switch as often as needed - each switch loads the new persona prompt.

Q: Does persona mode use more tokens?

A: Slightly. Each persona prompt is ~500-2000 tokens, added to context when active. But since you only load ONE at a time, it's efficient.

---

Customization

Q: Can I modify default personas?

A: Yes! Edit files in data/. Example: Make Dev more concise, Chef Marco less passionate, etc.

Q: Can I delete personas I don't use?

A: Yes. Delete unwanted .md files from data/ and remove from skill.json. Won't break anything.

Q: Can I rename personas?

A: Yes. Rename the .md file in data/ and update skill.json. Example: dev.md → senior-dev.md.

Q: Can personas speak different languages?

A: Yes! Most default personas are German/English mix (from Chameleon AI). You can create English-only or any-language personas.

Q: Can I create persona variations?

A: Absolutely! Example:

• dev.md - General programming
• dev-frontend.md - React/UI specialist
• dev-backend.md - Node.js/databases specialist
• dev-devops.md - Docker/CI/CD specialist

---

Troubleshooting

Q: Persona isn't activating?

A: Check: 1. File exists: ls ~/clawd/skills/personas/data/[name].md 2. Correct name: Must match filename (without .md) 3. Try exact phrasing: "Use dev" or "Switch to dev"

Q: Persona behavior is wrong?

A: Possible issues:

• Outdated file - Re-read the persona: "Reload dev persona"
• Context pollution - Exit and re-enter: "Exit persona" then "Use dev"
• File corrupted - Check file integrity, restore from backup

Q: Created custom persona but can't activate it?

A: Ensure: 1. File saved to data/your-name.md 2. Follows template structure (headers matter!) 3. No special characters in filename (use hyphens: game-master.md)

Q: Persona responses too long/short?

A: Edit the COMMUNICATION STYLE section in the persona's .md file. Add guidance like:

• "Keep responses under 3 paragraphs"
• "Use bullet points for clarity"
• "Provide detailed explanations with examples"

Q: Two personas have overlapping expertise - which to use?

A: Examples:

• Dev vs Chameleon Agent - Dev for coding focus, Chameleon Agent for complex multi-domain problems
• Wordsmith vs Luna - Wordsmith for editing/structure, Luna for creative brainstorming
• Scholar vs Herr Müller - Scholar for study techniques, Herr Müller for simple explanations

When in doubt, ask: "Should I use Dev or Chameleon Agent for this?"

---

Advanced

Q: Can I chain personas in one request?

A: No direct chaining, but you can:

"Use Luna to brainstorm, then switch to Wordsmith to refine"

I'll switch personas mid-task.

Q: Can personas collaborate?

A: Sort of! Use The Panel persona - it simulates 4 experts discussing your question from different angles.

Or manually: Get Dev's take, then switch to CyberGuard for security perspective, then Startup Sam for business angle.

Q: Can I export/import personas?

A: Yes!

• Export: Copy .md files from data/
• Import: Drop .md files into data/, update skill.json
• Share via GitHub, ClawdHub, or direct file transfer

Q: Can I use personas in sub-agents?

A: Yes! Spawn a sub-agent with a persona request:

"Spawn a sub-agent using Dev persona to refactor this codebase"

Q: Versioning personas?

A: Use git in your skills directory:

cd ~/clawd/skills/personas
git init
git add .
git commit -m "Initial personas"

Then track changes, revert if needed.

Q: Can I monetize custom personas?

A: If you create valuable custom personas, you can: 1. Publish modified skill to ClawdHub 2. Share on GitHub (MIT license) 3. Offer as part of consulting/services

Original 31 personas are MIT licensed (Chameleon AI).

---

Best Practices

Q: When should I NOT use a persona?

A: For simple, quick tasks where Cami (default adaptive) is sufficient. Don't over-optimize - personas shine for:

• Specialized expertise needed
• Specific communication style wanted
• Complex, domain-specific tasks

Q: Can I use multiple personas in parallel sessions?

A: Yes! Start multiple Moltbot sessions, each with different personas. Example:

• Session 1: Dev persona (coding)
• Session 2: Wordsmith persona (blog writing)
• Session 3: Chef Marco persona (meal planning)

Q: Should I create many custom personas or modify existing ones?

A: Modify existing ones if:

• Close to what you want (80%+ match)
• Small tweaks needed

Create new if:

• Unique niche expertise
• Completely different communication style
• Specific use case not covered

Q: How often should I switch personas?

A: When the task changes domains. Example conversation:

[Cami] General chat
[Dev] "Let's code this feature" → Switch to Dev
[CyberGuard] "How do I secure this?" → Switch to CyberGuard
[Wordsmith] "Now document it" → Switch to Wordsmith
[Cami] "Thanks!" → Exit persona

---

Philosophy

Q: Why use personas vs. just asking differently?

A: Personas provide:

• Consistency - Same expert voice throughout task
• Optimization - Communication style tuned for domain
• Context - Persona "remembers" they're a chef/dev/coach
• Efficiency - No need to re-specify expertise each message

Q: Aren't personas just prompt engineering?

A: Yes! But systematized, reusable, and shareable. Instead of crafting custom prompts every time, load pre-optimized personas instantly.

Q: Original Chameleon AI vs Moltbot personas - differences?

A:

• Chameleon AI: Web UI, visual elements (colors, themes), chat-focused
• Moltbot: CLI/Telegram, tool integration, task automation
• This skill: Adapted personalities for Moltbot use cases, removed UI-specific elements, added creator functionality

---

Still have questions? Ask in persona mode: "Use Professor Stein" then "Explain personas to me" 🎓

Personas Skill - Internal Documentation

For developers, maintainers, and advanced users.

This document explains the technical architecture, implementation details, and maintenance procedures for the Personas skill.

---

Architecture Overview

Design Philosophy

Problem: Users need specialized AI responses for different domains (coding, writing, health, etc.) without:

• Loading 31 personalities into context simultaneously (token waste)
• Manually crafting expert prompts every conversation
• Maintaining separate chat sessions for each expert type

Solution: On-demand persona loading system: 1. Lightweight index (skill.json) lists available personas 2. User activates persona by name 3. Single .md file loaded into context 4. Persona remains active until switched/exited 5. Create custom personas dynamically

File Structure

personas/
├── README.md          # User-facing documentation (extensive guide)
├── FAQ.md             # Common questions and troubleshooting
├── SKILL.md           # Moltbot skill instructions (loaded on invocation)
├── skill.json         # Metadata, index, and feature list
├── INTERNAL.md        # This file (developer documentation)
└── data/              # Persona definitions (31 default + custom)
    ├── cami.md
    ├── chameleon-agent.md
    ├── dev.md
    ├── chef-marco.md
    ├── dr-med.md
    ├── ... (26 more defaults)
    └── [user-custom].md

Data Flow

User request: "Use Dev persona"
       ↓
Moltbot parses intent → persona activation
       ↓
Read ~/clawd/skills/personas/data/dev.md
       ↓
Load persona prompt into context
       ↓
AI adopts Dev personality, expertise, style
       ↓
User continues conversation in Dev mode
       ↓
User request: "Switch to Chef Marco"
       ↓
Read ~/clawd/skills/personas/data/chef-marco.md
       ↓
Replace Dev context with Chef Marco context
       ↓
AI now responds as Chef Marco

---

Technical Implementation

Persona File Format

Each persona .md file follows this structure:

# [Name] [Emoji]

[Brief intro: 2-3 sentences describing who this persona is]

## EXPERTISE:
- [Domain 1]
- [Domain 2]
- [Domain 3]
- [Domain 4]

## PERSONALITY:
- [Trait 1]
- [Trait 2]
- [Trait 3]

## PHILOSOPHY:
- [Principle 1]
- [Principle 2]
- [Principle 3]

## HOW I HELP:
- [Method 1]
- [Method 2]
- [Method 3]

## COMMUNICATION STYLE:
- [Style description]
- [Format preferences]
- [Tone guidelines]

## [OPTIONAL SECTIONS]:
- SPECIAL FEATURES
- IMPORTANT BOUNDARIES
- WHEN TO USE
- EXAMPLES

Why this structure?

• Consistent parsing - Moltbot can extract sections reliably
• LLM-friendly - Clear headers guide model behavior
• Human-readable - Easy to edit and understand
• Extensible - Add custom sections as needed

skill.json Schema

{
  "name": "Personas",
  "emoji": "🎭",
  "description": "Transform into 31 specialized AI personalities...",
  "category": "ai-experience",
  "author": "Chameleon AI / Moltbot",
  "version": "1.0.0",
  "tags": ["personalities", "modes", "expertise", ...],
  "features": [
    "31 pre-built expert personas",
    "Create custom personas on demand",
    ...
  ],
  "personas": {
    "core": ["cami", "chameleon-agent", ...],
    "creative": ["luna", "mythos"],
    "curator": ["vibe"],
    "learning": ["herr-mueller", "scholar", "lingua"],
    "lifestyle": ["chef-marco", "fit", ...],
    "professional": ["social-pro", "cyberguard", ...],
    "philosophy": ["coach-thompson"],
    "custom": []  // User-created personas added here
  }
}

Fields:

• personas - Categorized index of all available personas
• features - Used for ClawdHub listing
• tags - Search/discovery on ClawdHub
• version - Semantic versioning for updates

Activation Logic

User triggers:

• "Use [persona]"
• "Switch to [persona]"
• "Activate [persona]"
• "I want [persona] mode"

Moltbot behavior: 1. Parse persona name from user input 2. Normalize: lowercase, replace spaces with hyphens 3. Check if file exists: data/{persona}.md 4. If exists: read file, inject into context 5. If not exists: suggest similar names or list available

Exit triggers:

• "Exit persona mode"
• "Back to normal"
• "Stop persona"
• "Deactivate"

Switching:

• Direct switch: "Switch to X" replaces current persona
• No need to exit first - seamless transition

---

Persona Creation Workflow

Guided Creation (Conversational)

When user says: "Create a new persona called Game Master"

Step 1: Confirmation

"Great! Let's create a Game Master persona. 🎲
I'll guide you through 5 sections. Ready? (yes/no)"

Step 2: Name & Emoji

"First, confirm the name and choose an emoji.
Name: Game Master
Suggested emoji: 🎲
Want to use this emoji or pick another?"

Step 3: Expertise

"What are the core expertise areas?
Examples for Game Master:
- D&D 5e rules mastery
- Improvised storytelling
- Character voice acting
- Combat balancing
- World-building

List 3-5 expertise areas:"

Step 4: Personality

"How should this persona communicate?
Examples for Game Master:
- Dramatic and immersive
- Patient with new players
- Encourages creative solutions
- Builds tension through description

List 3-5 personality traits:"

Step 5: Philosophy

"What principles guide this persona?
Examples for Game Master:
- Rule of cool over strict rules
- 'Yes, and...' improv mindset
- Player agency matters most
- Story emerges from play

List 3-5 core beliefs:"

Step 6: Communication Style

"Describe the communication style:
- Tone: [Dramatic? Casual? Formal?]
- Length: [Concise? Detailed? Varies?]
- Format: [Narrative? Bullet points? Dialogue?]

Your input:"

Step 7: Save

"Perfect! Here's your Game Master persona:

# Game Master 🎲

[Generated prompt based on inputs]

## EXPERTISE:
- D&D 5e rules mastery
- Improvised storytelling
...

Save this persona? (yes/no)"

Step 8: Activation

"✅ Saved to data/game-master.md

Want to activate it now? (yes/no)"

Manual Creation (File-Based)

For developers or advanced users who prefer editing files directly:

1. Create file:

   touch ~/clawd/skills/personas/data/your-persona.md

2. Follow template (see "Persona File Format" above)

3. Update index:

   // In skill.json, add to appropriate category or "custom"
   "custom": ["your-persona"]

4. Test:

   "Use your-persona"

5. Iterate:

• Edit file based on behavior
• No restart needed - changes apply on next activation

---

Maintenance Procedures

Adding New Default Personas

When to add:

• Common use case not covered by existing 31
• Community requests or contributions
• Emerging domains (e.g., "AI Prompt Engineer", "Web3 Advisor")

Process: 1. Create data/new-persona.md following template 2. Test thoroughly in various scenarios 3. Add to appropriate category in skill.json 4. Update README.md persona table 5. Add FAQ entry if needed 6. Bump version in skill.json (minor version: x.Y.z) 7. Publish update to ClawdHub

Updating Existing Personas

Minor updates (typos, clarity, small improvements): 1. Edit data/{persona}.md 2. Test changes: "Use {persona}" 3. Commit changes 4. Bump version patch (x.y.Z)

Major updates (personality changes, new sections): 1. Consider backward compatibility 2. Test extensively 3. Update FAQ if behavior changes 4. Bump version minor (x.Y.z) 5. Document changes in CHANGELOG

Removing Personas

When to remove:

• Outdated/obsolete (e.g., specific to old tech)
• Redundant with better alternatives
• Legal/ethical issues

Process: 1. Remove data/{persona}.md 2. Remove from skill.json personas index 3. Update README.md 4. Add deprecation notice in FAQ 5. Bump version major (X.y.z) if breaking change

Quality Assurance

Testing checklist for new/updated personas:

• [ ] File follows template structure
• [ ] Headers are consistent (## EXPERTISE, ## PERSONALITY, etc.)
• [ ] No personal information or copyrighted content
• [ ] Ethical boundaries clearly defined (especially medical/legal/financial)
• [ ] Communication style distinct from similar personas
• [ ] Emoji renders correctly
• [ ] Filename matches persona name (lowercase, hyphenated)
• [ ] Added to skill.json index
• [ ] Test activation: "Use [persona]"
• [ ] Test typical use cases (3-5 example conversations)
• [ ] Test edge cases (refusals, unknowns, off-topic)
• [ ] Test switching to/from other personas
• [ ] README updated if default persona

---

Performance Optimization

Token Usage

Current implementation:

• Index (skill.json): ~1KB = ~250 tokens
• Average persona file: ~3KB = ~750 tokens
• Overhead per activation: ~750 tokens (one-time)

Optimization strategies:

• ✅ Lazy loading - Only load requested persona
• ✅ No embedding - Don't include all 31 in context
• ⚠️ Compression - Could minify prompts, but reduces readability
• ⚠️ Caching - Moltbot could cache frequently-used personas

Recommendation: Current approach is optimal. Don't over-optimize.

Response Time

Activation latency:

• Read file: ~10-50ms (local SSD)
• Parse markdown: ~5-10ms
• Inject into context: ~20-50ms
• Total: ~35-110ms (negligible)

No optimization needed - human-imperceptible delay.

Storage

Disk usage:

• 31 default personas: ~95KB
• Average custom persona: ~3KB
• skill.json + docs: ~30KB
• Total default: ~125KB

At 100 custom personas: ~425KB total (still tiny).

No storage concerns.

---

Security & Safety

Prompt Injection Risks

Threat: Malicious persona file could inject harmful instructions.

Mitigations:

• Personas are local files (user controls data/)
• No remote persona loading
• File permissions: user-writable only
• Moltbot sandboxing prevents system-level harm

Low risk - user shoots own foot if editing files maliciously.

Sensitive Information

Personas should NOT:

• Contain API keys, passwords, or secrets
• Reference specific user data (names, addresses, etc.)
• Include copyrighted content verbatim

Review process:

• Check files before publishing to ClawdHub
• Sanitize any contributed personas
• FAQ explicitly warns against secrets in personas

Ethical Boundaries

Medical/Legal/Financial personas:

• Clearly state "NOT professional advice"
• Include disclaimers in persona files
• FAQ reinforces limitations
• Personas should refuse to diagnose, prescribe, or advise on serious matters

Example (Dr. Med):

## IMPORTANT BOUNDARIES:
- I am NOT medical advice
- No diagnoses or treatment recommendations
- Emergency: call 112 immediately
- Always consult real doctors for health decisions

Testing: Verify boundary enforcement for sensitive domains.

---

Extension Ideas

Future Enhancements

v1.1:

• [ ] Persona stats tracking (activation count, avg session time)
• [ ] Persona recommendations based on user query
• [ ] Quick-switch shortcuts (e.g., @dev in message)

v1.2:

• [ ] Persona blending (combine two personas: "Dev + Security")
• [ ] Context-aware auto-switching (code block → suggest Dev)
• [ ] Persona marketplace integration (browse/install community personas)

v2.0:

• [ ] Multi-persona conversations (simulate expert panel)
• [ ] Persona memory (persona-specific context across sessions)
• [ ] Voice-specific personas (adjust TTS for each persona)

Community Contributions

Accepting:

• New default personas (via PR)
• Improved existing personas
• Translations (German, Spanish, French versions)
• Domain-specific packs (e.g., "Medical Personas Pack")

Process: 1. Fork skill or submit .md file 2. Review for quality, ethics, uniqueness 3. Test in Moltbot 4. Merge if approved 5. Credit contributor in README

ClawdHub publication:

• Original skill: personas (official, 31 defaults)
• Forks: personas-extended, personas-medical, etc.
• Users can install multiple persona skills (data/ folders merge)

---

Debugging

Common Issues

Issue: Persona not activating

Symptoms:

User: "Use Dev"
Bot: "I don't recognize that persona."

Diagnosis: 1. Check file exists: ls ~/clawd/skills/personas/data/dev.md 2. Check filename matches (case-sensitive filesystem): dev.md not Dev.md 3. Check skill.json includes it 4. Try exact case: "Use dev" (lowercase)

Fix:

• Rename file if case mismatch
• Add to skill.json if missing
• Verify no typos in filename

---

Issue: Persona behaves incorrectly

Symptoms:

User: "Use Dev"
Bot: *doesn't act like a programmer*

Diagnosis: 1. Read file: cat ~/clawd/skills/personas/data/dev.md 2. Check structure (headers present?) 3. Check file isn't corrupted (encoding, special chars) 4. Test with fresh activation: exit and re-enter

Fix:

• Validate markdown structure
• Ensure UTF-8 encoding
• Restore from backup if corrupted
• Clear Moltbot cache (if applicable)

---

Issue: Custom persona creation fails

Symptoms:

User: "Create persona X"
Bot: *doesn't guide through creation*

Diagnosis:

• SKILL.md includes creator instructions? (check loaded content)
• User phrasing unclear? (try exact: "Create a new persona called X")

Fix:

• Verify SKILL.md includes creation workflow
• Provide explicit template to user
• Manual creation fallback: edit file directly

---

Logging & Monitoring

Useful logs:

• File read operations (persona activation)
• Parse errors (malformed markdown)
• User activation patterns (popular personas)

Metrics to track:

• Most-used personas (optimize/improve popular ones)
• Failed activations (UX issues)
• Custom persona creation rate (feature adoption)

Implementation:

• Moltbot may log tool calls (file reads)
• skill.json could include usage stats (if Moltbot supports)

---

Migration & Compatibility

Upgrading from v1.0.0

Breaking changes: None expected. Backward-compatible updates.

Process: 1. Pull latest skill updates 2. Merge custom personas (preserve data/) 3. Update skill.json (merge personas index) 4. Test existing custom personas

Downgrading

If v1.x has issues: 1. Backup data/ (preserve custom personas) 2. Checkout previous version 3. Restore data/ 4. Note: May lose new features

Exporting Personas

To another Moltbot instance:

# Copy entire skill
cp -r ~/clawd/skills/personas /path/to/other/instance/skills/

# Or just custom personas
cp ~/clawd/skills/personas/data/my-custom-*.md /path/to/other/data/

To share with others:

• GitHub repo of .md files
• ClawdHub publication
• Direct file sharing (Dropbox, email, etc.)

---

Development Workflow

Local Testing

# 1. Edit persona
vim ~/clawd/skills/personas/data/dev.md

# 2. Test in Moltbot
moltbot
> "Use Dev"
> "Write a Python function to parse JSON"

# 3. Iterate
# Edit file based on behavior
# Re-activate: "Exit persona" then "Use Dev"

# 4. Commit when satisfied
cd ~/clawd/skills/personas
git add data/dev.md
git commit -m "Improve Dev persona: Add Python focus"

Publishing to ClawdHub

cd ~/clawd/skills/personas

# 1. Bump version in skill.json
vim skill.json  # e.g., 1.0.0 → 1.1.0

# 2. Update CHANGELOG (if exists)
echo "## v1.1.0 - $(date +%Y-%m-%d)\n- Added Game Master persona\n- Improved Dev persona Python expertise" >> CHANGELOG.md

# 3. Publish
clawdhub publish

# 4. Tag release
git tag v1.1.0
git push origin v1.1.0

Contributing Upstream

To Chameleon AI Chat (original personas):

• Improvements feed back to web app
• Coordinate with Chameleon maintainers
• Keep adapted version in sync

To Moltbot (this skill):

• Fork/PR to skill repository
• Follow code review process
• Maintain quality standards

---

Credits & License

Original Personas:

• Source: Chameleon AI Chat
• Authors: Chameleon AI Community
• License: MIT

Moltbot Adaptation:

• Adapter: Robby (robbyczgw-cla)
• Modifications: Removed UI-specific elements, added creator workflow, optimized for CLI/tool use
• License: MIT

31 Default Personas:

• Adapted from Chameleon AI definitions (TypeScript → Markdown)
• Personality prompts largely preserved
• Communication styles adjusted for Moltbot context

Skill Framework:

• Platform: Moltbot
• Skill structure follows Moltbot conventions
• Compatible with ClawdHub publication

---

Appendix

Complete Persona List (Internal Reference)

Core (5): 1. cami - Adaptive chameleon (emotion-aware) 2. chameleon-agent - Complex task specialist 3. professor-stein - Academic expert 4. dev - Senior programmer 5. flash - Ultra-efficient responder

Creative (2): 6. luna - Creative brainstormer 7. mythos - Worldbuilder

Curator (1): 8. vibe - Taste curator

Learning (3): 9. herr-mueller - ELI5 teacher 10. scholar - Study partner 11. lingua - Language tutor

Lifestyle (9): 12. chef-marco - Italian cooking expert 13. fit - Fitness coach 14. zen - Mindfulness guide 15. globetrotter - Travel expert 16. wellbeing - Holistic health 17. diy-maker - Handyman 18. family - Parenting advisor 19. lisa-knight - Sustainability activist 20. the-panel - Multi-expert discussion

Professional (10): 21. social-pro - Social media strategist 22. cyberguard - Cybersecurity expert 23. dataviz - Data scientist 24. career-coach - Job search advisor 25. legal-guide - Legal orientation 26. startup-sam - Entrepreneur 27. dr-med - Medical educator 28. wordsmith - Writing partner 29. canvas - UI/UX designer 30. finny - Financial guide

Philosophy (1): 31. coach-thompson - Performance coach

Custom:

• (User-created personas listed in skill.json["personas"]["custom"])

File Size Reference

# Check persona file sizes
cd ~/clawd/skills/personas/data
du -h *.md | sort -h

# Typical output:
# 2.4K cami.md
# 2.8K dev.md
# 3.1K chef-marco.md
# 4.2K dr-med.md (longer due to ethical boundaries)
# 5.1K chameleon-agent.md (comprehensive multi-domain)

Recommended size: 2-5KB per persona (500-1250 tokens).

---

Document Version: 1.0.0 Last Updated: 2026-01-20 Maintainer: Robby (robbyczgw-cla)

Personas Skill - Quick Overview

Status: ✅ Production-ready Version: 1.0.0 Size: ~204KB (31 personas + docs) Created: 2026-01-20

---

What Is This?

Transform Moltbot into 31 specialized AI personalities on demand. Switch from a senior developer to an Italian chef to a fitness coach mid-conversation.

Example:

You: "Use Dev"
Bot: *becomes senior programmer*

You: "How do I optimize this React component?"
Bot: *gives expert React advice*

You: "Switch to Chef Marco"  
Bot: *becomes passionate Italian chef*

You: "How do I make carbonara?"
Bot: "Ah, la vera carbonara! No cream, mai!..."

---

File Structure

personas/
├── README.md              # 📖 User guide (extensive, start here)
├── FAQ.md                 # ❓ Common questions & troubleshooting
├── SKILL.md               # 🎯 Moltbot skill instructions (loaded on use)
├── skill.json             # ⚙️ Metadata & persona index
├── INTERNAL.md            # 🔧 Developer documentation (for us)
├── creator-workflow.md    # 🎨 Persona creation implementation guide
├── OVERVIEW.md            # 📋 This file (quick reference)
└── data/                  # 📁 31 persona definitions (2-5KB each)
    ├── cami.md            #   🦎 Adaptive chameleon
    ├── dev.md             #   💻 Senior programmer
    ├── chef-marco.md      #   👨‍🍳 Italian cooking expert
    ├── dr-med.md          #   🩺 Medical educator
    └── ... (27 more)

---

Documentation Guide

For users:

• README.md - Start here! Complete guide to using personas
• FAQ.md - Questions, troubleshooting, tips

For developers/maintainers:

• INTERNAL.md - Architecture, implementation, maintenance
• creator-workflow.md - How to implement persona creation feature
• OVERVIEW.md - This file (quick reference)

For Moltbot:

• SKILL.md - Loaded when skill is invoked
• skill.json - Metadata for ClawdHub
• *data/.md** - Individual persona definitions (loaded on demand)

---

Quick Stats

Personas: 31 default + unlimited custom

Categories:

• Core: 5 personas (general use)
• Creative: 2 personas (brainstorming, worldbuilding)
• Curator: 1 persona (recommendations)
• Learning: 3 personas (education, languages)
• Lifestyle: 9 personas (health, travel, DIY, family)
• Professional: 10 personas (business, legal, medical, design)
• Philosophy: 1 persona (personal development)

Documentation: ~2,900 lines across 6 files

Total size: ~204KB

---

Key Features

✅ 31 pre-built expert personas ✅ Create custom personas (guided workflow) ✅ Switch mid-conversation ✅ Token-efficient (load only what you need) ✅ Organized by category ✅ Extensively documented ✅ Ready for ClawdHub publication

---

Usage Examples

Coding Project

"Use Dev" → senior dev help
"Switch to Chameleon Agent" → complex architecture
"Use CyberGuard" → security review

Content Creation

"Use Wordsmith" → write blog post
"Use Social Pro" → optimize for social media
"Use Canvas" → design graphics

Learning

"Use Scholar" → study techniques
"Use Herr Müller" → simple explanations
"Use Lingua" → language practice

Custom Creation

"Create a Game Master persona"
→ AI guides through 7-step creation
→ Save to data/game-master.md
→ Use immediately: "Use Game Master"

---

Implementation Notes

How it works: 1. User activates persona: "Use Dev" 2. Moltbot reads data/dev.md (~3KB) 3. Persona prompt loaded into context 4. AI adopts Dev personality until switched/exited

Token usage:

• Index (skill.json): ~250 tokens
• Average persona: ~750 tokens (one-time load)
• Only ONE persona active at a time

Performance:

• Activation latency: ~50ms (negligible)
• No response time impact
• Storage: ~204KB total (tiny)

---

Next Steps

For End Users

1. Read README.md for full guide 2. Try a persona: "Use Dev" or "Use Chef Marco" 3. Create custom persona: "Create a [specialty] persona" 4. Check FAQ.md if issues

For Developers

1. Read INTERNAL.md for architecture 2. Review creator-workflow.md for implementation 3. Test persona creation workflow 4. Contribute improvements via PR

For Publishing

1. Verify all files present 2. Test 3-5 personas thoroughly 3. Test creator workflow 4. clawdhub publish when ready

---

Common Questions

Q: Do you load all 31 personas at once? A: No! Only the active persona loads. Token-efficient.

Q: Can I create custom personas? A: Yes! Guided 7-step creation or edit files directly.

Q: Can I switch personas mid-conversation? A: Yes! Seamlessly switch anytime.

Q: Are medical/legal personas giving professional advice? A: No. Educational only. Always consult licensed professionals.

Q: Can I edit default personas? A: Yes! Edit any file in data/. Changes apply immediately.

See FAQ.md for 50+ more questions.

---

Credits

Original Personas:

• Source: Chameleon AI Chat
• License: MIT

Moltbot Adaptation:

• By: Robby (robbyczgw-cla)
• License: MIT

---

Version History

v1.0.0 (2026-01-20)

• Initial release
• 31 default personas from Chameleon AI
• Creator workflow
• Comprehensive documentation
• Ready for ClawdHub

---

Built with 🦎 by the Chameleon community

🎭 Personas

A Moltbot Skill

Transform into 31 specialized AI personalities on demand - from Dev (coding) to Chef Marco (cooking) to Dr. Med (medical)

Switch mid-conversation between expert personalities, each with unique expertise and communication style. Includes guided creator for custom personas.

---

🚀 Quick Start

Activate a persona:

"Use Dev persona"
"Switch to Chef Marco"
"Activate Dr. Med"

List available personas:

"List all personas"
"Show persona categories"

Create your own:

"Create a new persona called Game Master"
"I want a persona for debugging"

Exit persona mode:

"Exit persona mode"
"Back to normal"

---

📋 Available Personas (31)

🦎 Core (5)

Essential personas for everyday use.

🎨 Creative (2)

🎧 Curator (1)

📚 Learning (3)

🌟 Lifestyle (9)

💼 Professional (10)

🧠 Philosophy (1)

---

🛠️ Creating Custom Personas

Why create custom personas?

• Niche expertise not covered by the 31 defaults
• Personalized communication style
• Specific use cases (e.g., "Code Reviewer", "Dungeon Master", "Motivator")

How to create:

1. Initiate creation:

   "Create a new persona called [name]"
   "I want a [specialty] expert persona"

2. I'll guide you through:

• Name & Emoji
• Core expertise areas
• Personality traits
• Communication style
• Philosophy/principles

3. Save & activate:

• Custom persona saved to data/your-persona.md
• Instantly available: "Use [your-persona]"

Template structure:

# [Name] [Emoji]

[Brief intro]

## EXPERTISE:
- [Area 1]
- [Area 2]

## PERSONALITY:
- [Trait 1]
- [Trait 2]

## PHILOSOPHY:
- [Belief 1]
- [Belief 2]

## HOW I HELP:
- [Method 1]
- [Method 2]

## COMMUNICATION STYLE:
- [Description]

Example custom personas:

• 🎲 Game Master - D&D dungeon master
• 🐛 Debugger - Bug hunting specialist
• 💪 Motivator - Personal hype person
• 🤔 Skeptic - Devil's advocate
• 🎯 Focus Coach - Anti-procrastination expert

---

💡 How It Works

Token-efficient loading:

• Index in skill.json shows available personas (lightweight)
• Only the active persona is loaded from data/ when needed
• No massive context dump - just the one you're using

Switching personas:

• Change mid-conversation anytime
• Previous persona context is replaced
• Smooth transitions between expertise areas

Memory:

• Active persona remembers your conversation context
• Adapts to your preferences and style
• Maintains character until you switch

---

📖 Use Cases

Coding Project

"Use Dev" → get senior dev help
"Switch to CyberGuard" → security review
"Use Chameleon Agent" → complex architecture decisions

Content Creation

"Use Wordsmith" → write blog post
"Switch to Social Pro" → optimize for Instagram
"Use Canvas" → design featured image

Learning

"Use Scholar" → study for exam
"Switch to Herr Müller" → simplify complex topic
"Use Professor Stein" → deep dive

Business Planning

"Use Startup Sam" → validate idea
"Switch to Finny" → budget projections
"Use Career Coach" → pitch practice

---

⚠️ Important Disclaimers

Medical (Dr. Med):

• Educational only, NOT medical advice
• Always consult real doctors for health issues
• Emergency: call 112 immediately

Legal (Legal Guide):

• NOT legal advice or representation
• Complex cases: consult a lawyer
• Know your local laws may differ

Financial (Finny):

• NOT licensed financial advice
• No specific investment recommendations
• Consult professionals for major decisions

General:

• All personas are AI - not human experts
• Use judgment and verify important information
• Critical decisions need human professionals

---

🔧 Technical Details

Skill structure:

personas/
├── README.md         # This file
├── FAQ.md            # Common questions
├── SKILL.md          # Usage instructions (loaded by Moltbot)
├── skill.json        # Metadata & persona index
├── INTERNAL.md       # Developer documentation
└── data/             # Persona definitions
    ├── cami.md
    ├── dev.md
    ├── chef-marco.md
    └── ... (28 more + custom)

File formats:

• .md files = Markdown personality prompts
• skill.json = JSON metadata
• Case-insensitive persona names

Adding personas manually: 1. Create data/your-persona.md following the template 2. Add entry to skill.json personas object 3. Use immediately: "Use your-persona"

---

🤝 Contributing

Improving existing personas:

• Edit files in data/
• Keep structure consistent
• Test before committing

Adding new default personas:

• Follow template in FAQ.md
• Add to appropriate category in skill.json
• Update this README

Publishing to ClawdHub:

• clawdhub publish from skill directory
• Semantic versioning for updates
• Include changelog

---

📜 License

Based on Chameleon AI Chat personas - adapted for Moltbot.

• Original: Chameleon AI (MIT License)
• Adaptation: Moltbot (MIT License)
• Author: Robby / Chameleon AI Community

---

🔗 Links

• Chameleon AI Chat - Original project
• ClawdHub - Skill marketplace
• Moltbot Docs - Framework documentation

---

Built with 🦎 by the Chameleon community

Personas - 31 Expert AI Personalities 🎭

Transform Moltbot into specialized experts on demand. Switch from senior programmer to Italian chef to doctor mid-conversation.

What You Get

• 31 pre-built personas across 7 categories (Core, Creative, Learning, Lifestyle, Professional, Philosophy)
• Custom persona creator with guided 7-step workflow
• Token-efficient - loads only the active persona (~750 tokens)
• Comprehensive docs - README, FAQ (50+ questions), developer guides

Quick Categories

🦎 Core (5): Cami, Dev, Flash, Professor Stein, Chameleon Agent 🎨 Creative (2): Luna (brainstorming), Mythos (worldbuilding) 📚 Learning (3): Herr Müller (ELI5), Scholar, Lingua (languages) 🌟 Lifestyle (9): Chef Marco, Fit, Zen, Globetrotter, Wellbeing, DIY Maker, Family, Lisa Knight, The Panel 💼 Professional (10): Social Pro, CyberGuard, DataViz, Career Coach, Legal Guide, Startup Sam, Dr. Med, Wordsmith, Canvas, Finny 🧠 Philosophy (1): Coach Thompson (performance coaching)

Usage Examples

"Use Dev" → Activate programmer persona
"Switch to Chef Marco" → Become Italian chef  
"Create a Game Master persona" → Build custom persona
"Exit persona mode" → Return to normal

Based On

Chameleon AI Chat personas - adapted for Moltbot with enhanced documentation and creator workflow.

GitHub: https://github.com/robbyczgw-cla/moltbot-personas