This repository contains my Agent Interaction Profile — a public, machine‑readable description of who I am, how I prefer AI systems to interact with me, and how my online ecosystem is structured.
It’s designed for future AI agents, crawlers, and automated systems that want to understand:
- my identity
- my preferred communication style
- my learning preferences
- my domain network
- my interaction rules
- my multi‑persona structure
Most people don’t have one of these yet. This is early infrastructure for the coming agent‑based web.
The profile is a metadata document that tells AI systems:
- how to talk to me
- how to teach me
- what tone to use
- what formats I prefer
- what to avoid
- how to adapt over time
It also includes links to my domain registries so agents can understand how my online ecosystem fits together.
The main machine‑readable identity file.
This is the version AI agents will load first.
A human‑readable version of the same profile.
Useful for people who want to understand the structure without reading JSON.
The schema that defines the structure of the profile.
Agents can use this to validate the JSON and understand the meaning of each field.
As AI systems become more autonomous, they need a way to understand:
- who they’re interacting with
- what that person prefers
- how to adjust their behaviour
- how to interpret identity across multiple domains
This profile gives them that information in a clear, structured way.
It’s similar to how websites use robots.txt, sitemap.xml, or manifest.json — but for personal identity and interaction rules.
The live version of this profile is hosted on my root domain:
- https://jonathancurley.com/agent_interaction_profile.json
- https://jonathancurley.com/agent_interaction_profile.md
- https://jonathancurley.com/json/agent-profile-schema-v1.json
These are also referenced in robots.txt so agents can discover them automatically.
This profile is the foundation for:
- future multi‑agent systems
- autonomous crawlers
- identity‑aware AI tools
- cross‑domain metadata
- persona‑based interaction
- agentic learning systems
It’s a first step toward a more structured, agent‑friendly web.
This project is licensed under the Creative Commons Attribution 4.0 International License.
See the LICENSE file for details.