How to share Organization-Wide Core Memory across Repositories

Many engineering teams have baseline policies (coding style, security constraints, deployment protocols, API design rules) that apply across all repositories in an organization.

This cookbook shows you how to structure shared organization-wide core memory alongside project-specific memory.

Strategy 1: Shared Git Submodule / Template

Maintain canonical organization standards in a shared repo (e.g. github.com/my-org/engineering-standards):

1. Store shared guidelines in standards/core-org.md.
2. Include or reference it in each project's .memofs/memory/core.md:

# Project Core Memory

## Organization Standards (Inherited)
- Security: All public API endpoints require HMAC or JWT validation.
- Tooling: Default to TypeScript strict mode with pnpm package manager.
- Formatting: Follow Biome style rules.

## Project-Specific Rules
- Database: Primary database is PostgreSQL on AWS RDS.
- Auth: Multi-tenant auth with organization tenancy keys.

Strategy 2: MemoFS Cloud Shared Workspace

When using MemoFS Cloud, organization rules can be managed in an Organization Workspace:

1. Create an Organization Workspace in the MemoFS Cloud Dashboard.
2. Set the organization ID in your .memofs/config.json:

{
  "runtime": "hybrid",
  "workspaceId": "ws_engineering_org",
  "projectId": "proj_billing_service",
  "cloud": {
    "baseUrl": "https://api.memofs.dev/v1"
  }
}

3. MemoFS context queries will blend organization-level constraints with project-level memories seamlessly.

Best Practices

• Keep Core Memory Lean: Core memory is loaded on every agent session. Reserve core.md for non-negotiable rules; store detailed guides in notes.md.
• Use Classified Kinds: Tag organization rules with --kind constraint or --kind decision so they receive appropriate decay thresholds.

Related Resources

• Core Memory Concepts
• Configuration File Specs
• Team Memory Sync