Skip to content

Home

Jump to bottom
Paola Di Maio edited this page Feb 1, 2026 · 1 revision

🌐 Agentic AI Open Standards – Gap Analysis

Editor: @starborn
 Status: Community Discussion Draft
 Scope: Web & Open Standards for Agent Interoperability

🧭 Why This Exists

The web platform has mature open standards for:

  • Identity

  • Data exchange

  • Knowledge representation

  • Service interfaces

However, it does not yet have a shared semantic and behavioral framework for autonomous AI agents acting on the web.

As AI systems evolve from passive tools into goal-directed, decision-making agents, the lack of interoperability standards is becoming a critical gap.

This repository is an effort to map those gaps and help the open standards community coordinate work toward agent-native web standards.

🧠 What We Mean by “Agentic AI”

In this work, an agent is a software system that can:

  • Pursue goals

  • Make decisions with some autonomy

  • Interact with humans, services, data, and other agents

  • Potentially learn or adapt over time

This goes beyond APIs or chatbots — toward web-native autonomous actors.

📊 What’s in This Repo

Resource | Description -- | -- Standards Gap Analysis (Spreadsheet) | A structured map of agent capabilities vs existing and missing standards Technical Explainer | A narrative overview of the interoperability problem and why new standards are needed

Together, these documents aim to answer:

Where do current web standards fall short for autonomous agents?

🧩 Key Gap Areas Identified So Far

Some of the most important missing layers include:

  • Goal specification — No standard way to express machine-readable goals

  • Planning semantics — No interoperable model for plans, strategies, or constraints

  • Agent identity — No identity profile for autonomous non-human actors

  • Delegation & authority — No model for what an agent is allowed to decide or act on

  • Memory models — No shared representation of agent memory structures

  • Policy & governance — No standard way to attach operational or legal constraints

  • Agent-to-agent communication semantics — Protocols exist, but not shared meaning

🏗 Why This Belongs in Open Standards

Without open standards:

  • Agents become platform-bound silos

  • Interoperability depends on private APIs

  • Governance and safety become non-transparent

  • The web risks losing its open, decentralized nature

Open standards can help ensure that agent ecosystems are:

  • Interoperable

  • Auditable

  • Governable

  • Innovation-friendly

🤝 How to Contribute

This is an early-stage community bootstrap effort.

We welcome:

  • Missing standards we should reference

  • Corrections to gap assessments

  • Proposals for new ontology areas

  • Links to relevant W3C, IETF, ISO, IEEE, or industry work

  • Use cases that stress-test the model

Please open an Issue or Pull Request.

📜 License

Suggested: CC-BY 4.0 (for documents)
 Standards work benefits from maximum reuse.

Clone this wiki locally