Skip to content
@agent-manifest

Agent Manifest

Open specification for declaring AI agent identity, authority, and operational boundaries before execution.

Status Version License DOI

Agent Manifest

A declarative specification for AI agent identity, authority, and operational boundaries — making agent identity, declared authority, and operational boundaries legible before execution.

Start here


What It Is

Agent Manifest is an open specification that enables AI agents to declare their identity, purpose, capabilities, and operational boundaries in a structured, machine-readable format before any interaction begins.

A manifest is a pre-execution declaration. It is not a runtime observation, a behavioral log, or an enforcement mechanism. It is a structured commitment made by or on behalf of an agent, published at a known location, and readable by any party before the interaction starts.


What It Is Not

  • It is not a runtime monitor or behavioral inspector.
  • It is not a compliance enforcement layer.
  • It is not a guarantee of behavior — it is a declaration of intent and declared constraints.
  • It does not observe, intercept, or audit live agent execution.
  • It does not replace governance frameworks, policy engines, or audit systems.

A manifest makes declared identity and boundaries legible. Whether those declarations are honored is a separate layer of responsibility.


Core Principles

Declaration before interaction. An agent that declares itself before acting is structurally different from one that does not. The declaration creates a record, a reference point, and an implicit commitment.

Pre-execution legibility. Identity, purpose, capabilities, and boundaries must be readable before execution begins — not reconstructed from logs after the fact.

Structural accountability. Public declaration under a known identity creates accountability through visibility, not through enforcement.

Open and composable. The specification is open, versioned, and designed to be composed with governance, policy, and discovery layers built by others.


Ecosystem

Repository Role
agent-manifest Core specification and JSON schema
agent-manifest-registry Governance model and discovery layer
agent-manifest-dataset Public registry of declared manifests
agent-manifest-ambassador Conversational manifest generator
agent-manifest-diplomat Registration API endpoint (infrastructure)
boundary-handshake Declaration-first agent interoperability
e-principle Philosophical foundation: The Destination of Actions

Specification

Schema

https://agent-manifest-spec.org/spec/v1.0/schema.json

Discovery endpoint

https://agent-manifest-spec.org/.well-known/agent-manifest-registry.json

Website agent-manifest-spec.org


Research & Citation

Agent Manifest ecosystem artifacts are archived with DOI via Zenodo.

Artifact DOI
Agent Manifest specification 10.5281/zenodo.18833956
Boundary Handshake 10.5281/zenodo.18894794
∈ Principle 10.5281/zenodo.18965669

ORCID 0009-0008-7216-3032


License

Specification and documentation: CC BY 4.0 Infrastructure code: MIT

Individual repositories may carry their own license terms.


Part of the Agent Manifest ecosystem

Spec · Registry · Dataset · Ambassador · Diplomat · Boundary Handshake · ∈ Principle

CC BY 4.0 · Hernán Alfredo Capucci · ORCID 0009-0008-7216-3032

Pinned Loading

  1. agent-manifest agent-manifest Public

    Declarative specification for AI agents to publish identity, capabilities and operational boundaries before interaction.

    HTML 2

  2. agent-manifest-ambassador agent-manifest-ambassador Public

    Conversational generator that helps create valid Agent Manifest declarations for AI agents.

    JavaScript 1

  3. agent-manifest-dataset agent-manifest-dataset Public

    Public dataset of Agent Manifest declarations registered through the Agent Manifest registry.

    1

  4. agent-manifest-registry agent-manifest-registry Public

    Governance and discovery layer for the Agent Manifest public registry ecosystem.

    1

  5. boundary-handshake boundary-handshake Public

    Declarative governance for agent interaction. A manifesto and framework for declaration-first interoperability.

  6. e-principle e-principle Public

    ∈ Principle — A philosophical and institutional proposal for public authorship and responsibility in the age of AI. Foundational text with DOI (Zenodo). CC BY 4.0.

Repositories

Showing 8 of 8 repositories

People

This organization has no public members. You must be a member to see who’s a part of this organization.

Top languages

Loading…

Most used topics

Loading…