Dandori
The organizational outer harness for AI coding agents.
Your engineers already have great agents — Claude Code, Codex, Cursor, Cline. What’s missing is the organizational layer around them: shared context, cost control, approval gates, quality measurement, and audit trail.
That’s Dandori. One platform, two audiences — engineers use it every day, leadership sees through it every week.
What is an Outer Harness? — The concept behind Dandori: why process and data matter more than the agent you’re using.
The 5 pillars
| # | Pillar | What it solves |
|---|---|---|
| 1 | Cost Attribution | “$180K bill — where did it go?” |
| 2 | Knowledge Flow | “Agent violated policy — engineer forgot to paste compliance doc” |
| 3 | Task Tracking | “Who approved the migration that broke prod?” |
| 4 | Quality Gates | “Is agent output getting better or worse?” |
| 5 | Audit & Analytics | “Show me every AI-generated change in Q3” |
Two principles: process-centric (human and agent follow the same pipeline) · data-driven (every operation produces structured data).
Explore
| Page | What you’ll find |
|---|---|
| Outer Harness | The concept: inner vs outer, 5 pillars, two principles |
| Dandori Overview | What Dandori is + the 13 features under the 5 pillars + differentiators |
| Workflows | 8 iconic scenarios (4 leadership + 4 engineer) with sequence diagrams showing component interactions |
| Architecture | Tech stack, system diagram, adapter layer, deployment |
| Roadmap | Milestones from team pilot to full vision |
| Data Inventory | What data Dandori touches — for security review |
Read next
Outer Harness → The concept behind Dandori
Bring governance to your agents.
We’re onboarding design-partner organizations running 1,000+ engineers.