Paste sales call transcripts into the repo and get structured analyses that score which prospects show real buyer demand versus polite interest.
Build a documented ideal customer profile backed by actual conversation evidence rather than assumptions, then use it to brief your AI writing tools.
Set up MCP server connections to your CRM and run AI-assisted outreach drafts that trace back to specific demand signals from past calls.
Use Claude Code slash commands to run a PULL Framework analysis on a new transcript and instantly see how the prospect scores on all four conditions.
Requires an AI coding assistant such as Claude Code or Cursor, CRM integration is optional but needs MCP server configuration.
GTM Context OS is a git repository template designed to help sales and marketing teams build and run their go-to-market work using AI tools. The idea is that the repository itself becomes a shared memory system: AI coding assistants read it for context and write back to it with results, so the system gets smarter over each sales conversation analyzed. The template has two modes. Context mode is for building understanding. You feed in sales call transcripts, and the system produces structured analyses that score each call for real buyer demand. After several analyses, patterns across calls become visible: who is actually ready to buy, what language buyers use, what objections keep appearing. This feeds into documents that define your ideal customer profile and your competitive positioning. Operational mode is for acting on that understanding. You connect your CRM and other sales tools through MCP server configurations, write scripts that pull transcripts or enrich contact records, draft outreach sequences, and track campaign results. The two modes share the same repository so findings from conversations feed directly into campaign planning. The methodology is built around what the project calls the PULL Framework, credited to Rob Snyder. PULL stands for four conditions: the prospect has a Project on their list, that project is Unavoidable right now, they are Looking at options, and those options are Lacking. The template is structured so that every piece of messaging or outreach must trace back to documented demand evidence, not assumptions. The repository works with any AI coding assistant that reads an AGENTS.md file. Claude Code, Cursor, GitHub Copilot, Windsurf, and others are all supported through their respective config files, which all point back to the same AGENTS.md instructions. Claude Code users also get a set of slash commands for common tasks like running a PULL analysis or checking campaign status. You clone the repository, disconnect it from the template origin so it becomes fully your own, and open it in your AI editor. The template publishes updates as patterns with explanations rather than force-merging changes, so adopting improvements is always a deliberate choice.
← thesnappingdog on gitmyhub — every repo by this author, as a profile.
Verify against the repo before relying on details.