
gtm-engineer-starter
Start here: Jordan Crawford's GTM engineering method as Claude Code / Codex skills — foundation wiki, account dossiers, won/lost/10X, evidence packets scored…
Install with your AI
Paste into Claude Code, Cursor, or any agent — it reads the repo and wires the tool into your project.
Install and set up gtm-engineer-starter (git-clone project) into my current project. Found on https://claudeers.com/gtm-engineer-starter Repo: https://github.com/SantaJordan/gtm-engineer-starter Homepage/docs: — Detected install method: git-clone → git clone https://github.com/SantaJordan/gtm-engineer-starter Category: data. Platforms: cli, api. Read the repo's README for exact setup and env vars, then install it and wire it into my project. Claudeers Health Verdict: active; community-verified: false. Confirm the source before running anything.
git clone https://github.com/SantaJordan/gtm-engineer-starter
// compatibility
| Platforms | cli, api |
|---|---|
| Operating systems | — |
| AI compatibility | claude |
| License | MIT |
| Pricing | open-source |
| Language | Python |
GTM Engineer Starter
A starter kit for go-to-market (GTM) engineering with Claude Code or Codex: find out why your best customers bought, turn that into rules, and score a prospect list against those rules with evidence you can show.
It is the free, open-source version of the method Jordan Crawford teaches at Blueprint. It ships six skills, a project template, and one fictional worked example that runs end to end with no API keys.
New here? Open START-HERE.md. It is the method in order, for people and for AI agents.
What you can do with it
- Build account dossiers. Join your CRM, calls and billing into one timeline per account, with the source of every fact.
- Label Won / Lost / 10X. Work backward from which customers stayed, which left, and which brought in at least 10 times the median customer's revenue, to what was true before they bought.
- Write your ICP as rules. Inclusions and exclusions, each traced to a real account. ICP means ideal customer profile: the kind of company that buys and stays.
- Score a list with evidence. One evidence packet per prospect, judged by Jev, a calibrated model that answers with probabilities. Unsure rows go to a human review queue instead of being guessed.
- Write a PVP message. A message so specific the buyer would pay to read it, with every claim tied to a source.
Quick start
git clone https://github.com/SantaJordan/gtm-engineer-starter ~/gtm
cd ~/gtm
python3 -m pip install -r requirements.txt
make check
make check runs every test suite, an offline smoke test, and an offline replay
of the whole method on a fictional company. It needs no API keys and spends
nothing. You should see 6/6 steps passed. and 13/13 replay steps passed.
What is in the box
| Folder | What it is |
|---|---|
START-HERE.md | The method, step by step, with commands and decision rules |
foundation/ | The wiki pattern, how to connect your data, how to keep it in a private GitHub repo, and a project template |
skills/dossier-data-auditor/ | Checks your data before a dossier build: joins, domains, revenue columns |
skills/dossier-builder/ | Builds per-account dossiers across every system |
skills/win-loss-rewind/ | Won / Lost / 10X: finds the signals that separate good customers from bad |
skills/crawford/ | Builds evidence packets for a list (the packet builder from the Crawford research agent) |
skills/calibrated-judge/ | Scores packets with Jev; writes scores and a human review queue |
skills/wiki-lint/ | Weekly health check for your wiki |
examples/ | A fictional company, Northwind, run end to end, plus the acceptance check |
TOOLS.md | Tools and vendors, with links |
NOT-INCLUDED.md | What the paid version adds |
llms.txt | A file index for AI agents |
Using the skills
Claude Code: each folder in skills/ has a SKILL.md with a name and
description. From your project folder run make install-skills, which links every
folder into the project's .claude/skills/ (or copy them there, or into
~/.claude/skills/ for every project). It prints six SKILL.md paths when it
worked. Then ask for a skill by name, for example /dossier-data-auditor. Keep
crawford/ and calibrated-judge/ side by side: the judge finds the packet
builder there. START-HERE.md walks through this just before Step 3.
Codex or another agent: nothing to install. Point it at START-HERE.md and
llms.txt, and start a step with "Read skills/<name>/SKILL.md and follow it".
Every SKILL.md is plain markdown, and every script runs from the command line.
Keys
Nothing is needed for make check. For real work, copy .env.example to .env
and fill in only what you use. The two that matter most:
- a Jev key for scoring:
TYPESAFE_API_KEY(preferred) orOPENROUTER_API_KEY; ANTHROPIC_API_KEYfor Won / Lost / 10X.
.env is gitignored. Do not commit keys.
Your data stays yours
This repo holds only fictional data. When you add your own, keep the copy in a
private repository. .gitignore already keeps .env, raw exports
(wiki/raw/, private/), customer CSVs at the root and generated runs out of
git. GitHub, your AI assistant's provider and any API you configure still receive
what you send them. See foundation/push-to-github.md.
License
MIT. Copyright 2026 Jordan Crawford / Blueprint. See LICENSE.
The wiki pattern builds on Andrej Karpathy's LLM Wiki and Jacob Dietl's gtm-context-os-quickstart. Jev is a model from TypeSafe, used through its API; no TypeSafe code is included.
// faq
What is gtm-engineer-starter?
Start here: Jordan Crawford's GTM engineering method as Claude Code / Codex skills — foundation wiki, account dossiers, won/lost/10X, evidence packets scored with Jev.. It is open-source on GitHub.
Is gtm-engineer-starter free to use?
gtm-engineer-starter is open-source under the MIT license, so it is free to use.
What category does gtm-engineer-starter belong to?
gtm-engineer-starter is listed under data in the Claudeers registry of Claude-compatible tools.
// embed badge
[](https://claudeers.com/gtm-engineer-starter)
// retro hit counter
[](https://claudeers.com/gtm-engineer-starter)
// reviews
// guestbook
// related in Data & Analytics
Marketing skills for Claude Code and AI agents. CRO, copywriting, SEO, analytics, and growth engineering.
Wrap Antigravity, ChatGPT Codex, Claude Code, Grok Build as an OpenAI/Gemini/Claude/Codex compatible API service, allowing you to enjoy the free Gemini 3.1 P…
Apache Airflow - A platform to programmatically author, schedule, and monitor workflows
CLI tool for configuring and monitoring Claude Code