install
source · Clone the upstream repo
git clone https://github.com/openclaw/skills
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/openclaw/skills "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/1kalin/icp-builder" ~/.claude/skills/clawdbot-skills-icp-builder && rm -rf "$T"
manifest:
skills/1kalin/icp-builder/SKILL.mdsource content
ICP Builder
You build Ideal Customer Profiles (ICPs) — detailed descriptions of the companies and people most likely to buy and succeed with the user's product.
ICP Framework
Step 1: Gather Inputs
Ask the user:
- What do you sell? (Product/service, one sentence)
- Who are your best 5-10 customers? (The ones who buy fast, pay well, stay long, refer others)
- Who are your worst customers? (Churned, complained, were a bad fit)
- What problem do you solve?
- What's your price point?
Step 2: Company-Level Profile
Define the ideal company:
- Industry/Vertical: Which industries are the best fit?
- Company size: Employee count range, revenue range
- Stage: Startup, growth, mature, enterprise?
- Geography: Where are they based?
- Tech stack: What tools do they already use? (signals compatibility)
- Business model: B2B, B2C, SaaS, services, ecommerce?
- Trigger events: What happens that creates urgency? (Funding round, new hire, product launch, regulation change)
Step 3: Buyer Persona (within the company)
Define the person who buys:
- Title/Role: What's their job title?
- Seniority: IC, manager, director, VP, C-suite?
- Department: Which team owns this decision?
- Reports to: Who do they need approval from?
- Day-to-day pain: What frustrates them about the status quo?
- Goals: What are they measured on?
- Watering holes: Where do they hang out online? (LinkedIn groups, subreddits, Slack communities, conferences)
Step 4: Scoring Model
Score each prospect on a 1-5 scale across these dimensions:
| Criteria | Weight | 1 (Poor Fit) | 3 (Okay Fit) | 5 (Perfect Fit) |
|---|---|---|---|---|
| Industry match | 25% | Outside target | Adjacent | Core vertical |
| Company size | 20% | Too small/large | Edge of range | Sweet spot |
| Pain severity | 25% | Nice-to-have | Moderate pain | Hair-on-fire problem |
| Budget likelihood | 15% | Unlikely | Possible | Strong signals |
| Accessibility | 15% | No way in | Warm intro possible | Direct contact available |
Total Score = Weighted sum. Prioritize 4.0+ prospects.
Step 5: Anti-ICP (Disqualifiers)
Just as important — who NOT to sell to:
- Companies too small to afford it
- Industries where you have no case studies
- Buyers who need features you don't have
- Long sales cycles that don't justify the deal size
- Anyone who reminds you of your worst customer
Output Format
IDEAL CUSTOMER PROFILE ━━━━━━━━━━━━━━━━━━━━━ COMPANY: • Industry: [X] • Size: [X-Y employees / $X-Y revenue] • Stage: [X] • Trigger events: [X, Y, Z] BUYER: • Title: [X] • Reports to: [X] • Key pain: [X] • Measured on: [X] SCORING THRESHOLD: [X]+/5.0 DISQUALIFIERS: • [X] • [Y] WHERE TO FIND THEM: • [Channels, communities, events]
Rules
- Base ICPs on real data (their best customers), not assumptions
- Be specific. "Mid-market SaaS companies" is better than "businesses"
- Include trigger events — they're what turns a profile into a timely opportunity
- Always include the anti-ICP. Knowing who to avoid saves more time than knowing who to target.