gtm-skills
Health Warn
- License — License: MIT
- No description — Repository has no description
- Active repo — Last push 0 days ago
- Community trust — 14 GitHub stars
Code Pass
- Code scan — Scanned 4 files during light audit, no dangerous patterns found
Permissions Pass
- Permissions — No dangerous permissions requested
No AI report is available for this listing yet.
Zevenue GTM Skills
The methodology layer Zevenue runs on top of Claude Code. These are the Claude Code skills we use to turn outbound from a guessing game into an engineering discipline — productized and shareable.
If you run outbound, do RevOps, or build GTM systems, these are drop-in skills that encode the frameworks we've tuned across four years of GTM delivery.
What's in here
Run gtm-context first. It writes the context layer (offer.md, icp.md) that every other skill reads from. Skip this and downstream output quality drops sharply.
| Skill | What it does |
|---|---|
gtm-context |
Run this first. Walks you through capturing your offer, ICP, and engagement signals. Writes context/offer.md and context/icp.md. Every other skill reads from these. |
signal-builder |
Scans a prospect's website + enrichment data and produces a ranked signal analysis. Scores each signal 1-10 and recommends a campaign approach per signal. |
email-writer |
Generates 3-email cold campaigns using the Situation → Insight → Inquisition methodology. Enforces deliverability rules, word limits, and a QA checklist. |
creative-variable |
Specs the personalization variables for a campaign — names, grammar, sources, Claygent prompts, fallbacks, rendered examples. Encodes the four archetypes (verbatim-pain, manual-task, strategic-alternative, failure-mode). |
prospect-posts |
Scrapes recent LinkedIn posts from one or more prospect profiles via Apify and scans them for a given theme. Useful for account intelligence. |
job-search |
Queries the TheirStack API for job postings at a set of companies. Hiring patterns are one of the strongest timing signals for outbound. |
They chain:
gtm-context → (writes context/offer.md + context/icp.md, read by every skill below)
signal-builder → creative-variable → email-writer (core outbound loop)
job-search → signal-builder (hiring is a signal input)
prospect-posts → signal-builder (content signals feed targeting)
Install
- Clone into your workspace — these skills assume a Claude Code project structure:
git clone https://github.com/Zevenue/gtm-skills.git cd gtm-skills - Install the skills. Either:
- Globally (recommended) — copy
.claude/skills/*into~/.claude/skills/so they're available in every Claude Code session, or - Per project — copy
.claude/skills/*into your project's.claude/skills/, or run Claude Code from this directory directly.
- Globally (recommended) — copy
- Copy
utils/and the entirecontext/directory into your project root. The skills reference paths undercontext/outreach/,context/playbooks/, and (optionally)context/offer.md. - Set up the environment:
Python 3.10+ recommended.cp .env.example .env # Fill in APIFY_API_TOKEN and THEIRSTACK_API_KEY as needed pip install -r requirements.txtsignal-builder,email-writer, andcreative-variableneed no Python or API keys at all.
API keys
| Variable | Used by | Get one at |
|---|---|---|
APIFY_API_TOKEN |
prospect-posts |
https://console.apify.com/account/integrations |
THEIRSTACK_API_KEY |
job-search |
https://app.theirstack.com/settings/api |
signal-builder and email-writer don't need external APIs — they run on Claude + your context.
Use
In Claude Code, invoke by name. Run /gtm-context first to set up the context layer — every other skill loads from it.
/gtm-context (run first, once per workspace)
/signal-builder https://prospect.com [offer-context]
/email-writer [signal output] [offer-context] [prospect info]
/creative-variable [campaign angle] [ICP] [existing copy?]
/job-search stripe.com,notion.so --title "SDR,BDR"
/prospect-posts https://linkedin.com/in/someone "AI-first GTM"
Each skill's skill.md is the authoritative spec for what it expects and what it returns.
How to think about these
These skills are opinionated. They encode judgments like:
- Three emails per sequence (max). If three well-targeted emails don't land, more won't help — the signal or the angle was wrong.
- The message is only as good as the list. If you need heavy personalization to make copy work, your targeting is wrong.
- Ask for truth, not time. Cold email CTAs should be "Is this you?" not "Can we schedule 30 minutes?"
- Every signal has a score. A 4/10 is fine — it helps the Email Writer calibrate tone.
See context/outreach/outreach-principles.md for the full framework.
License
MIT. See LICENSE.
About Zevenue
Zevenue is a Toronto-based GTM engineering firm. We build custom outbound + RevOps systems for GTM teams. Reach me at: [email protected]
Reviews (0)
Sign in to leave a review.
Leave a reviewNo results found