Agent Identity

An AI agent’s identity is defined by its actions, context, and constraints—not merely by a name or static scope.

​

Why We Need More Than RBAC

In today’s AI-driven world, agents perform critical tasks autonomously. Beyond asking whether an agent can perform an action, we must define when, why, and under what context it should—and should not—act. Traditional Role-Based Access Control (RBAC) employs static, binary scopes (e.g., gmail.read, slack.post) that quickly become unwieldy across multiple roles and environments.

Scenario: Rahul Parundekar uses a single inbox (rahul@aihero.studio) to:

• Oversee investor communications as CEO
• Address on-call alerts as DevOps Lead
• Manage RSVPs as Community Organizer
• Analyze portfolio performance as Investor

A blanket gmail.read permission would let a DevOps agent inspect payroll emails or allow an Investor bot to archive deployment alerts—outcomes that violate principle of least privilege.

​

A Contextual Ontology: Roles, Profiles, Agents

To resolve these limitations, we propose three foundational concepts:

1. Role Defines a set of context-aware, semantic permission clauses within an Organization.
2. Profile Binds a Person to a specific Role within an Organization.
3. Agent A software principal that assumes a Profile, acting on behalf of the associated Person.

​

1. Role

A Role lives inside an organization and comprises:

• organization: Identifier (e.g., AI Hero Inc. or org-12345)

• title: Role name (e.g., CEO, DevOps)

• capabilities: A list of high-level policy clauses (RFC‑style):

  • MUST: Obligations
  • SHOULD: Recommendations
  • MUST NOT: Prohibitions
  • MAY or CAN: Optional capabilities

# DevOps @ AI Hero Inc.
id: role-devops-aihero
organization: AI Hero Inc.
title: DevOps
capabilities:
  - CAN read infrastructure alert emails.
  - MUST verify sender domains match approved analytics providers.
  - SHOULD tag processed alerts with label "DevOps".
  - MUST NOT access financial or HR-related messages.

​

2. Profile

A Profile associates a Person with a Role. Profiles contain no additional instructions and represent the static binding between a person and a role. Agents may assume any existing profile.

# Rahul's CEO profile
id: profile-ceo-aihero
person: Rahul Parundekar
role: role-ceo-aihero

​

Persona

A Persona extends a Profile by including task-specific instructions. Personas encapsulate both the role and the operational guidance an agent should follow.

1. Newsletter Manager Persona

id: persona-newsletter-manager-aihero
role: role-ceo-aihero
instructions:
  - Fetch and tag weekly newsletter emails.
  - Send a summary to internal-comm@aihero.studio every Monday at 10:00 PT.
  - Archive processed newsletters into folder "Newsletters".

2. Board Interaction Persona

id: persona-board-interaction-aihero
role: role-ceo-aihero
instructions:
  - Prioritize emails containing "Board" or "Investor" keywords.
  - Schedule follow-up meetings in Calendar when requested.
  - Log meeting minutes to the #board channel on Slack.

3. Team Morale Persona

id: persona-team-morale-aihero
role: role-ceo-aihero
instructions:
  - Check the #team-happiness Slack channel daily.
  - Send positive feedback or shout-outs to team members.
  - Flag any morale issues by creating a ticket in the HR system.

​

ProfileConstraint

A ProfileConstraint defines the subset of profiles or personas an agent is authorized to assume.

id: constraint-rahul-assumable-profiles
agent: chatgpt-01
allowed_profiles:
  - persona-newsletter-manager-aihero
  - persona-board-interaction-aihero
  - persona-team-morale-aihero

​

3. Agent

An Agent is a programmatic identity that selects one Profile before acting:

agent: chatgpt-01
profile: profile-ceo-aihero

When an agent executes, its capabilities and constraints derive directly from its current profile.

​

Example: Fine-Grained Gmail Permissions

Model Rahul’s CEO profile with contextual Gmail access:

# CEO @ AI Hero Inc.
id: role-ceo-aihero
organization: AI Hero Inc.
title: CEO
capabilities:
  - MUST read emails containing investor or board-related keywords.
  - SHOULD archive newsletters into folder "CEO-News".
  - MAY send emails from rahul@aihero.studio.
  - MUST log sent-message IDs to the audit service.
  - MUST NOT operate outside 09:00–17:00 PT.

​

Advantages

• Least Privilege by Context: Agents operate only within their profile’s scope and temporal constraints.
• Expressive Policy DSL: Free-text, RFC-like clauses capture real-world nuances.
• Unified Model: No separate contract layer; Roles encapsulate both scope and policy.

​

Future Directions

In an upcoming post, we will explore how to integrate this context-rich identity model with standard authentication frameworks (such as OAuth 2.1) to enforce these roles and profiles in practice. By combining our ontological approach with robust authentication tokens and protocols, we aim to complete the picture for secure, end-to-end agent authorization. Stay tuned!

​

Get Connected, Share, and Other Socials