> ## Documentation Index
> Fetch the complete documentation index at: https://handlebar.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Core Concepts

> The elements you'll work with using Handlebar AI governance

These concepts are deliberately simple. Most complexity in agent systems comes from unclear boundaries. We're strict about those boundaries.

## Agents

An **Agent** is any system that can:

* Reason
* Decide
* Invoke tools

Handlebar does not care how the agent is implemented (LLM, workflow engine, hand-written logic). We treat the agent as an external actor whose behaviour must be governed.

**Agents are not trusted by default.**

## Tools

A **Tool** is any capability an agent can invoke:

* API calls
* Database queries
* File access
* Emails, messages, jobs, side effects

**Tools are where agents touch the real world. That makes them the highest-risk surface.**

In Handlebar, tools carry metadata:

* What kind of data they touch
* Who they affect
* Whether they're reversible
* Where outputs go

This metadata is critical for rule evaluation.

## Tool Calls

A **Tool Call** is a specific invocation of a tool during a run.

It includes:

* Inputs
* Context
* Timing
* Outcome (if allowed)

**Tool calls are the primary decision points in the system.** Rules are evaluated before a tool call is executed.

## Rules

A **Rule** is a deterministic check evaluated against context.

**Given the same inputs, a rule will always return the same result.**

Rules can inspect:

* Agent identity
* Run context
* Tool metadata
* Tool call parameters
* Historical signals

**Rules are versioned. Old decisions can be replayed exactly.**

## Actions

An **Action** is the consequence of a rule evaluation.

**Actions are hard, not advisory.** Examples:

* Allow
* Block
* Modify inputs
* Require approval
* Kill a run
* Lockdown

**If an action blocks a tool call, the tool does not run. There is no bypass.**

## Determinism

**The same inputs produce the same decisions.**

This enables:

* Replay
* Debugging
* Audits
* Trust

**If a decision can't be explained deterministically, it doesn't belong in the control path.**

## Why This Matters

In regulated environments, - "this is unlikely to happen" - isn't good enough.

You need to be able to say:

* This agent could not access this data under these conditions
* This action was blocked for this reason
* Here is the exact rule version that made that decision

That's the bar for Handlebar.