10 open-source repos. 127 files. The infrastructure layer for how brands, products, and design will work when everything becomes agentic.
Experimental Project | Every repo is exploratory, not production-ready. Read the DISCLAIMER and DYOR files in each repo before using anything. No warranties. No professional advice. Use at your own risk.
Built by Dragoon.
AI agents are about to represent brands, negotiate on behalf of users, make design decisions at scale, and interact with products autonomously. The entire infrastructure for this doesn't exist yet.
Brand guides are PDFs. User preferences are siloed. Trust signals are unreadable legal docs. Design systems assume static interfaces. None of it was built for a world where agents act on your behalf.
This collection fixes that. Each repo solves one piece of the puzzle. Together they form the missing infrastructure layer for the agentic era.
These three repos are the base. Everything else builds on top of them.
| # | Repo | What It Does |
|---|---|---|
| 01 | Brand Genome Project | Maps brand behavior as structured decision DNA that agents can read and execute. Not visual systems. Behavioral systems. |
| 02 | Anti-Style Guide | Documents what your brand would NEVER do. Constraints are more useful than instructions for generative AI. |
| 03 | Product Soul Contracts | A machine-readable declaration of what a product is, what it promises, and what it will never do. The trust handshake agents need. |
These repos give agents the knowledge and context they need to make good decisions.
| # | Repo | What It Does |
|---|---|---|
| 04 | Brand Memory Architecture | How brands store decisions, outcomes, and lessons in a format agents can query and learn from. Institutional memory for agentic orgs. |
| 05 | Preference Graph Standard | A universal format for user preferences that travel across products and agents. Your taste, style, and values become portable. |
| 06 | Design Drift Tracker | Monitors how AI-generated design decisions accumulate and drift from original brand intent over time. Git for brand coherence. |
These repos design the interaction patterns and testing frameworks for living with agents.
| # | Repo | What It Does |
|---|---|---|
| 07 | Ritual Library | The emotional choreography of human-agent interaction. Not UI patterns. Rituals. The pause before action. The recovery after mistakes. |
| 08 | Phantom Personas | AI-powered synthetic users with memory, grudges, and boredom that stress-test your brand over simulated time. |
| 09 | Agent Clash Testing | What happens when a user's agent negotiates with a brand's agent? Interaction logs, patterns, failure modes, and design recommendations. |
| 10 | Negotiable Interface | A design system where nothing is fixed. Every element is a negotiated agreement between user and product agents. |
Every repo ships with:
- Schemas defining the data structures (JSON Schema)
- Templates you can fill in for your own brand or product (YAML)
- Examples showing the frameworks applied to real-world scenarios
- Working code (Python and Node.js) for validation, simulation, and scoring
- Documentation explaining the thinking, the process, and the integration path
- Legal protection (DISCLAIMER, DYOR, LEGAL, LICENSE) in every repo
Total: 127 files across 10 repositories.
If you're a brand or marketing leader: Start with Brand Genome Project and Anti-Style Guide. These immediately change how you think about brand documentation for an AI-first world.
If you're a product designer: Start with Ritual Library and Negotiable Interface. These open up entirely new design disciplines.
If you're a developer building with AI agents: Start with Product Soul Contracts and Preference Graph Standard. These are the protocols your agents need.
If you're a founder: Start with the Foundation Layer in order. Then use Phantom Personas to stress-test everything before you ship.
Every one of these repos exists because of a simple observation: the tools and standards we use for brands, products, and design were built for a world where humans do everything. That world is ending.
When agents represent your brand, they need behavioral DNA, not logo files. When agents encounter your product, they need trust contracts, not terms of service. When agents make design decisions, someone needs to track the drift. When agents negotiate on behalf of users, someone needs to design the rules of engagement.
None of these systems existed. Now they do.
New here? Read the Use Cases & How-To Guide for real-world scenarios and step-by-step instructions for every repo.
Every repo accepts contributions. If you find edge cases, improve a schema, add a template for a new industry, or build an integration, submit a PR. See the CONTRIBUTING.md in each repo for guidelines.
This entire collection is experimental. It explores emerging concepts at the intersection of AI agents, branding, product design, and automated systems. The field is moving fast.
Before using anything:
- Read DISCLAIMER.md for full legal disclaimers
- Read DYOR.md for guidance on doing your own research
- Read LEGAL.md for liability and indemnification terms
Do your own research. Test everything. Get professional advice when it matters. Use your own judgment.
MIT License. See LICENSE in each repo for full terms.
Built by Dragoon. Shared freely. Build something with it.