Your rules live in CLAUDE.md here, .cursorrules there and AGENTS.md somewhere else. Blueprints keeps every rule that shapes your AI in one versioned, team-shared library, imported straight from GitHub and organized per tool.
CLAUDE.md in one repo, .cursorrules in another, AGENTS.md somewhere else. No single source of truth for how your AI should behave.
You fix a convention in one place and forget the other four, and each tool slowly behaves a little differently.
New chat, new tool, new teammate: the same standards, explained from scratch, again.
Connect a repository through the GitHub App and VibeXP imports the AI config you already keep there, classifying each file by the tool and kind it belongs to. Re-imports skip what is already there.
# your-org/payments-api CLAUDE.md Claude Code root file AGENTS.md Codex root file .claude/agents/reviewer.md Claude Code sub-agent
Every blueprint is typed for the tool it configures, so your AI reaches the right rules for the right tool. Your agent pulls them over MCP; nothing is written back to your repo.
vibexp_io_get_resource( resource_type: "blueprint", slug: "claude-md" )
A blueprint is a stored, versioned Markdown document holding the rules and config that govern an AI tool: coding standards, instruction files such as CLAUDE.md and AGENTS.md, and tool config such as sub-agents, skills and slash-commands. A prompt is the task you ask for; a blueprint is the rule it always follows.
In the docs ↗Yes. Connect your repository through the team GitHub App and import its AI-config files as per-tool blueprints. Only Markdown files are imported, each is classified by tool and kind, and re-imports skip blueprints that already exist.
In the docs ↗No. Import is one-directional, from GitHub into VibeXP. Your AI reads blueprints over MCP, and you can copy any of them anywhere you need it.
In the docs ↗Stars are how the next person tired of re-explaining their project to an AI finds this. It is the whole of the distribution: no ad budget, no sales team, no hosted plan to upsell.
Free, open source and self-hosted. One command to run it, one to connect your AI tools, and everything else in the docs.