Shared Context Engineering

A methodology that treats shared, versioned context as a first-class engineering artifact. Your AI tools pull in the context they need, when they need it.

What is SCE?

SCE is a simple idea: keep a context/ folder in your repository that documents your architecture, patterns, and decisions. AI tools read this folder automatically, so they understand your project without you repeating yourself.

your-project/
context/
overview.md
architecture.md
patterns.md
decisions.md
src/
package.json

To get started, you give your AI agent a system prompt that tells it how to read and use your shared context. From there, every conversation begins with your project already understood.

Why it matters

Faster onboarding

New team members and AI tools understand your codebase immediately.

Consistent output

AI follows your patterns and conventions because it knows what they are.

Preserved decisions

Architectural choices and their reasoning stay documented and accessible.

Works with any AI coding assistant

Just plain markdown files in your repo. No proprietary formats, no vendor lock-in.

Tooling & MetricsClosed Beta

SCE works on its own with just markdown files. But if you want deeper insights, we offer optional tooling to help you understand how your team works with AI and track engineering performance.

Currently supports Claude Code, OpenCode, and Cursor with more integrations coming soon.

We are working to open source everything. Interested? Get in touch.

AI Usage Analytics

Understand how your team adopts AI tools. Track usage patterns, measure time saved, identify what works, and optimize your AI investment.

Engineering Metrics

Measure what matters for your team. Cycle time, deployment frequency, PR throughput, code review time, and DORA metrics to drive continuous improvement.

Context Health Monitoring

Keep your context fresh and accurate. Get alerts when documentation drifts from reality, track coverage, and ensure your AI always has up-to-date information.

Join the Beta

We are working to open source everything.

Want to participate in shaping the tooling and get early access? Reach out and help us build it.

Docs

Why SCE Exists

Learn the motivation behind Shared Context Engineering and why teams need it as AI adoption scales.

Read the guide

Ready to get started?

Set up SCE in your project in about 10 minutes.

Building on great ideas. SCE continues and improves on the ideas from Effective Context Engineering for AI Agents by Anthropic and LODE by @fjzeit. We took what works and shaped it into a practical, team-first methodology with tooling to match.

Stay in the loop. Get notified when our tooling hits open beta and goes open source.