Client Data Directory Structure

Standard root: /home/claude/ge-bootstrap/ge-ops/clients/

Each client gets a directory using a URL-safe slug of their company name.

ge-ops/clients/
├── {client-slug}/
│   ├── intake/              ← Dima: client uploads, raw materials
│   ├── scope/               ← Aimée: scope documents, research
│   ├── design/              ← Alexander: design assets, mockups
│   ├── specs/               ← Anna: formal specifications
│   ├── packages/            ← Faye: work package definitions
│   ├── development/         ← Team: working code, branches
│   ├── testing/             ← Marije/Judith: test results, reports
│   ├── delivery/            ← Leon/Rutger: deployment configs
│   └── communications/      ← Margot: client correspondence log

Rules: - Client data NEVER leaves its slug directory - Each stage adds to the client folder, never overwrites previous stages - Agents read upstream folders (Aimée reads intake/), write to their own - Tenant isolation: no agent reads across client directories without explicit authorization