AI Systems · Software and Coding

Agents and Tool Use

An AI agent is a software system that pursues a scoped objective through a sequence of steps, sometimes using models, retrieval, tools, rules, and human input. A tool extends what the system can observe or do. Tool availability does not provide intent, permission, or proof that the result is correct.

Sources & review

How Atlas supports this record.

Atlas uses original Kinesema wording. These public sources support narrow parts of the record; they are not endorsements or proof of claims outside the roles shown below.

Review
Human reviewed
AI assistance
Yes—human reviewed
Last reviewed
2026-08-23

AI Agent Standards Initiative

National Institute of Standards and Technology

Used here for
NIST describes agent systems capable of autonomous actions on behalf of users and identifies interoperability, security, authentication, identity, and human-agent interaction as active concerns.
Source use
Paraphrased from source
Source type
official emerging standards and research initiative overview

Lessons Learned from the Consortium: Tool Use in Agent Systems

National Institute of Standards and Technology and AI Safety Institute Consortium

Used here for
Current agent systems may embed general-purpose models in wider software scaffolding that uses tools; tool use can be examined by function, permissions, environment, risk, reliability, monitoring, and autonomy.
Source use
Paraphrased from source
Source type
official workshop findings and preliminary taxonomy discussion

Tools — Model Context Protocol Specification 2025-11-25

Model Context Protocol project

Used here for
MCP provides one concrete interface in which servers expose named tools with schemas, clients invoke them, and human denial and untrusted metadata remain explicit.
Source use
Paraphrased from source
Source type
official stable protocol specification

Artificial Intelligence Risk Management Framework (AI RMF 1.0)

National Institute of Standards and Technology

Used here for
Wider AI-system and deployment context, lifecycle evaluation, human and organizational roles, and the need to test and monitor system behavior and interpret outputs in context.
Source use
Paraphrased from source
Source type
official government risk-management framework; version 1.0; published 2023

This is a concise first-release record, not a complete course. Examples, terminology, and reviewed connections can grow through later public projections.

Connected concepts

These are reviewed Atlas relationships, not suggestions inferred from page order or visual proximity. Open a boundary only when you want the extra detail.

Connects to

Retrieval and Evaluation

An agent may use retrieval to obtain material and evaluation to compare behavior or outcomes with a stated goal, policy, or expectation.

Connection boundary

Retrieving material or passing an evaluation does not grant permission to act, validate every tool result, or replace human review.

Connected from

Computing and AI Systems

An agent is a wider software-system role that may coordinate models, retrieval, tools, rules, and human input within a computing environment.

Connection boundary

This relationship does not make every AI system an agent or give an agent authority merely because a capability is available.