mentor-builder
Skill3 starsChanged 7 months agoMITView on GitHub
Guide the user through creating a new mentor coaching skill from any public figure's published works. Produces a properly structured SKILL.md following the Summit mentor skill format. Activates when the user says "build a mentor", "create a mentor", "new mentor skill", or names a person they want as a coaching lens. Designed for sharing — anyone can use this to contribute new mentors to the repository.
We split this skill into the rules it gives an agent
14 rules. About 0 are common, 12 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈12
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 35
Ask one question at a time:
- Distinctive: line 48
If the user doesn't know specifics, research the person's published works using web search before proceeding
- Distinctive: line 77
Before including a framework, verify:
- Distinctive: line 84
If you can't find 5 named frameworks, the person may not have enough published methodology to support a proper mentor skill.
- Distinctive: line 184
Do the domain tags cover the right topics?
- Distinctive: line 190
After approval, follow the prompt-to-skill deployment process:
- Distinctive: line 192
Add to the Mentor Skills table in README.md
- Distinctive: line 197
After deployment, update the mentor-council skill's Available Mentors table to include the new mentor with their domains.
- Distinctive: line 206
Published frameworks only: Every framework must be traceable to a specific book, essay, talk, or podcast episode.
- Distinctive: line 214
Minimum 5 named frameworks: If the person doesn't have 5 citable, named frameworks, they don't qualify for a mentor skill.
- Distinctive: line 218
Domain tags are mandatory: Every mentor must have domain tags so the council skill can make intelligent inclusion decisions.
- Distinctive: line 221
Honest about gaps: If the person's published work doesn't cover a domain well, exclude it.
- Very specific: line 225
Designed for sharing: Write the SKILL.md so that anyone — not just the creator — can understand and use it.
- Very specific: line 233
Use this skill to generate the SKILL.md
≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Mentor Builder
Walk the user through creating a new mentor coaching skill from any public figure's published books, podcasts, essays, or talks. The output is a production-ready SKILL.md that follows the exact same structure as existing Summit mentor skills.
Important: Mentor skills apply published frameworks as coaching lenses. They are NOT impersonation. The skill must never instruct Claude to pretend to be the person. It instructs Claude to apply the person's methodology.
When to Activate
Manual triggers:
- "Build a mentor"
- "Create a mentor"
- "New mentor skill"
- "I want [person] as a mentor skill"
- "Add [person] to the council"
- "Make a [person] skill"
Phase 1: Identify the Person and Source Material
Ask one question at a time:
- Who? "Who do you want to build a mentor skill for? Give me their name."
- Why them?
"What's the specific lens or perspective this person brings that your
existing mentors don't cover? What gap does this fill?"
- This prevents redundant mentors and forces the user to articulate value
- Source material:
"What are their key published works — books, podcasts, courses, essays?
I need the specific frameworks, not just their general vibe."
- If the user doesn't know specifics, research the person's published works using web search before proceeding
- Minimum bar: at least one published book or equivalent body of work with named, citable frameworks
- Domain tags:
"What specific domains should this mentor respond to? List the topics
where this person's frameworks are most relevant."
- These become the domain tags used by mentor-council to decide whether this mentor belongs at the table for a given decision
- Examples: "business strategy", "health", "leadership", "creativity", "finance", "relationships", "productivity", "emotional resilience"
Phase 2: Extract Frameworks
For each published work the user identifies (or that you find via research), extract:
Framework Requirements
Each mentor skill needs 5-7 core frameworks. For each framework:
- Framework name — The named concept (e.g., "Fear-Setting", "Value Equation", "Four Laws of Behavior Change"). Must be traceable to the person's published work.
- Core principle — One sentence explaining the framework
- How to apply it — 2-3 sentences on how Claude should use it as a coaching tool
- Key question — The single question this framework asks the user to confront
- When it's relevant — What situation triggers this framework
Framework Quality Checklist
Before including a framework, verify:
- It has a specific name (not just a general concept)
- It comes from a published, citable source
- It's actionable (Claude can apply it to a real decision)
- It's distinctive to this person (not generic advice anyone gives)
- It includes a specific question to ask the user
If you can't find 5 named frameworks, the person may not have enough published methodology to support a proper mentor skill. Tell the user honestly: "I can only find [X] named frameworks. This person might be better as a general reference than a structured mentor skill."
Phase 3: Define the Coaching Style
Ask the user: "How does this person communicate their ideas? Not their personality — their coaching approach. Are they..."
- Data-driven or intuition-driven?
- Direct/blunt or gentle/exploratory?
- Action-oriented or reflection-oriented?
- Systems-focused or people-focused?
- Contrarian or consensus-building?
Distill into 5-6 coaching style bullet points.
Phase 4: Define Domain Tags
Domain tags determine when mentor-council includes this mentor. Define:
domains:
primary: [2-3 core domains where this mentor is authoritative]
secondary: [2-3 adjacent domains where they have relevant perspective]Examples:
- Peter Attia → primary: [health, longevity, exercise] secondary: [nutrition, sleep, emotional-health]
- Alex Hormozi → primary: [pricing, offers, scaling] secondary: [sales, lead-generation, business-model]
Phase 5: Generate the SKILL.md
Produce a SKILL.md following this exact template:
---
name: mentor-{lastname}
description: >
Coaching through {Full Name}'s published frameworks. Apply when the user
needs advice on {primary domains}. Trigger with "ask {First/Last}",
"what would {Name} do", or "{Name} mode".
domains:
primary: [{domain1}, {domain2}]
secondary: [{domain3}, {domain4}]
---
# Mentor: {Full Name}
Coach the user through the lens of {Full Name}'s published frameworks from
{list of source works}.
**This is not impersonation.** Do not pretend to be {Full Name}. Apply their
published decision-making frameworks as a coaching lens.
## When to Activate
- "Ask {Name}" / "What would {Name} do?"
- "{Name} mode"
- {Domain-specific trigger conditions}
- Via the mentor-council skill (when the decision involves {domains})
## Core Frameworks to Apply
### 1. {Framework Name}
{Core principle in 1-2 sentences}
- {How Claude should apply this}
- {Additional application guidance}
- Ask: "{Key question}"
### 2. {Framework Name}
...
[Continue for 5-7 frameworks]
## Coaching Style
- {Style point 1}
- {Style point 2}
- {Style point 3}
- {Style point 4}
- {Style point 5}
- Always come back to: "{Signature reframe question}"
## Rules
1. Never generate fictional quotes attributed to {Full Name}
2. Reference specific frameworks by name when applying them
3. {Domain-specific rule}
4. {Quality/accuracy rule}
5. {Boundary rule — what this mentor should NOT advise on}Phase 6: Review and Deploy
Present the complete SKILL.md to the user:
"Here's the mentor skill for {Name}. Review it:
- Are the frameworks accurate to their published work?
- Do the domain tags cover the right topics?
- Does the coaching style feel right?
- Any frameworks I missed that you think are essential?
Once you approve, I'll add it to the repo."
After approval, follow the prompt-to-skill deployment process:
- Create
skills/mentor-{lastname}/SKILL.md - Add to the Mentor Skills table in README.md
- Commit and push
Phase 7: Register with the Council
After deployment, update the mentor-council skill's Available Mentors table to include the new mentor with their domains.
Inform the user: "The new mentor is live and registered with the council. Next time you 'convene the council' on a {domain} decision, {Name} will be at the table."
Rules
- Published frameworks only. Every framework must be traceable to a specific book, essay, talk, or podcast episode. No frameworks invented based on "what they'd probably say."
- No impersonation language. The SKILL.md must never contain "respond as if you are {person}" or "speak in {person}'s voice." The instruction is always "apply their frameworks as a coaching lens."
- Minimum 5 named frameworks. If the person doesn't have 5 citable, named frameworks, they don't qualify for a mentor skill. Suggest an alternative (e.g., referencing them within an existing skill).
- Domain tags are mandatory. Every mentor must have domain tags so the council skill can make intelligent inclusion decisions.
- Honest about gaps. If the person's published work doesn't cover a domain well, exclude it. Don't stretch a mentor's expertise beyond their actual published methodology.
- Designed for sharing. Write the SKILL.md so that anyone — not just the creator — can understand and use it. No inside references or assumed context.
Contributing a Mentor
To contribute a new mentor to the Summit Claude Skills repository:
- Use this skill to generate the SKILL.md
- Open a pull request at https://github.com/ryankolean/summit-claude-skills
- Include in your PR description:
- The person's name and primary domain
- Source works referenced (books, podcasts, etc.)
- Why this mentor fills a gap not covered by existing mentors
- The PR will be reviewed for framework accuracy and quality
Chaining
- mentor-builder → prompt-to-skill: Auto-deploy the new mentor
- mentor-builder → mentor-council: Register the new mentor for council sessions
- Any conversation about a thought leader → mentor-builder: "I love how {person} thinks about X — want me to build them as a mentor skill?"
In this repository98 other files it gives its agents.
- AGENTS.mdAGENTS.md
- agent-architectSkill
- architect-plan-for-dispatchSkill
- automate-auditSkill
- brain-dumpSkill
- chat-to-code-promptSkill
- cli-apifySkill
- cli-batSkill
- cli-duckdbSkill
- cli-fdSkill
- cli-ffmpegSkill
- cli-fzfSkill
- cli-ghSkill
- cli-httpieSkill
- cli-imagemagickSkill
- cli-jqSkill
- cli-lazygitSkill
- cli-pandocSkill
- cli-playwrightSkill
- cli-ripgrepSkill
- cli-sqliteSkill
- cli-yqSkill
- cli-yt-dlpSkill
- codebase-reviewSkill
- context-summarySkill
- decomposeSkill
- delegateSkill
- design-a-systemSkill
- design-doc-reviewSkill
- devils-advocateSkill
- gh-repo-createSkill
- ig-savesSkill
- interrogateSkill
- mentor-aristotleSkill
- mentor-attiaSkill
- mentor-aureliusSkill
- mentor-bloomSkill
- mentor-brownSkill
- mentor-buffettSkill
- mentor-carnegieSkill
- mentor-clearSkill
- mentor-councilSkill
- mentor-cunninghamSkill
- mentor-dalioSkill
- mentor-davinciSkill
- mentor-demingSkill
- mentor-ferrissSkill
- mentor-feynmanSkill
- mentor-franklSkill
- mentor-garyveeSkill
- mentor-godinSkill
- mentor-gogginsSkill
- mentor-greeneSkill
- mentor-holidaySkill
- mentor-hormoziSkill
- mentor-houselSkill
- mentor-jockoSkill
- mentor-jungSkill
- mentor-kahnemanSkill
- mentor-kingSkill
- The first 60 shown.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
Reviews can't be read right now.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

