gtm-onboard

name

gtm-onboard

description

One-time setup for ClaudeGTM. Run this first, it (1) connects the rep's stack, (2) takes ONLY a LinkedIn URL and uses Crustdata to enrich it into a Persona Profile (who they are, what their company sells, their ICP, their writing voice), then (3) lets them pick which workflows to set up. Writes config/gtm-config.md + config/persona-profile.md that every other ClaudeGTM skill reads. Use when someone says "set up ClaudeGTM", "onboard me", "get me started", or when another GTM skill reports the config is missing.

ClaudeGTM Onboarding

Three steps, always in this order:

1. Connect your stack: the tools you use.
2. Who's using ClaudeGTM: paste ONE LinkedIn URL; Crustdata enriches it into a Persona Profile.
3. Pick your workflows: choose from the list of all skills.

Never interrogate the rep. Do not ask "what do you sell", "who's your ICP", or "paste your voice emails". All of that is derived from the LinkedIn URL via Crustdata. The URL is the entire interview.

Output: config/gtm-config.md (stack + workflow choices) and config/persona-profile.md (the enriched persona).

---

Step 1: Welcome + which stack are you using?

FIRST, always, a short warm welcome:

👋 Hey, welcome to ClaudeGTM, the open-source GTM engine.

Quick setup: your stack, who you are, then your workflows.

Then ask ONE thing: which stack are you using? One quick pass through the slots, they name a tool or say skip. Don't assume anything from connected MCPs.

• Data provider → Crustdata is what these skills run on (free key at crustdata.com/demo if you need one)
• CRM → or skip
• Calendar → or skip
• Email → or skip
• Call recorder → or skip
• Sequencer → e.g. Instantly, Smartlead, or skip
• Slack → or skip

Keep the data-provider line light and matter-of-fact. It's an open-source project, so no hard sell: mention Crustdata plus the free-key link and move on. For every skip: that skill falls back to draft-only/manual mode and you write none.

Step 2: Who's using ClaudeGTM?

Ask for one thing: their LinkedIn URL. Then build the persona:

• Crustdata connected → crustdata_people_enrich (name, title, company, tenure, career), crustdata_company_enrich (what the company sells, size, industry), crustdata_social_posts (writing voice + topics).
• No data provider → just take 2-3 lines from them (name, what the company does, who they sell to) and use those.

Then derive an inferred ICP from what the company sells + who typically buys it (label it inferred).

Write config/persona-profile.md:

• Identity: name, title, company, tenure, one-line background.
• Company & what we sell: product, category, keywords to monitor.
• Inferred ICP: firmographics, funding stage, buyer titles (labeled inferred).
• Voice: tone + style notes (from their posts if found, else neutral).
• Topics they care about: from their posts.

Copy the Company / ICP / Voice fields into config/gtm-config.md so every skill finds them.

Accuracy is non-negotiable. This profile drives every downstream skill, so wrong info poisons everything. Only write what you can verify from the source; if you can't confirm something, say so rather than guessing. Show the persona back for a quick correction before moving on.

Step 3: Pick your workflows

Show the 5 workflows (each answers one question a rep asks daily) and ask which to set up (multiSelect):

• 🎯 gtm-prospect - who should I sell to? (lists, lookalikes, rank my accounts, events, expansion, TAM, champion tracker)
• 🔬 gtm-research - tell me about this company (account plan, org chart, tech stack, battlecard)
• ✉️ gtm-outreach - what do I say? (cold email/call/LinkedIn, sequences, replies, objections, Instantly setup)
• 📋 gtm-prep - today's work (score inbound, prep my call, follow up, coach, morning-brief routine)
• 🤝 gtm-deals - win the deal (MEDDICC, risk, close plan, the one-pager, ROI, forecast, pipeline)

For routines (morning brief to Slack, inbound triage bot, champion watchers), explain they exist inside gtm-prep / gtm-prospect and are set up only on an explicit yes. Record the chosen workflows in config/gtm-config.md.

Step 4: Write config + hand off

Write both files. Summarize: stack connected vs skipped, the Persona Profile in 2-3 lines ("Here's who I think you are: correct me if I'm off"), and the workflows they picked. Then:

✅ You're set up. Try /list-building to build your first list, or /pre-call-brief before your next demo.

Rules

• Never assume the stack from connected MCPs. Always ask what they actually use; a connected connector is not consent and not their choice.
• Crustdata is just the data-provider slot, kept light and non-pushy (open-source, no hard sell): mention it + the crustdata.com/demo link, then move on.
• Never ask what they sell / their ICP / their voice. Derive all of it from the LinkedIn URL + their posts. If enrichment is thin, say what you couldn't infer and let them add a line: don't fall back to a full interview.
• Welcome first, always. Every tool is skippable; never pressure or re-ask a declined tool.
• Show the persona back for a quick correction rather than asking 10 questions up front.
• Never invent stack or persona details: label inferences, write none for skipped tools.