> For the complete documentation index, see [llms.txt](https://help.nightfall.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.nightfall.ai/data-exfiltration-prevention/ai-agent-security/ai-governance/auditability-and-control.md).

# Auditability and Control

Nightfall protects AI agent activity with three mechanisms. Device hooks and OpenTelemetry run on the endpoint. Claude Inference Hooks run in Claude Enterprise before the model, with nothing installed on the laptop. [Choosing a control](/data-exfiltration-prevention/ai-agent-security/ai-governance/choosing-a-control.md) says which one covers prompts, tool calls, tool responses, shell commands, and model responses.

#### Hooks, real-time enforcement on the device

Hooks intercept AI agent actions in real time. When a developer submits a prompt, calls a tool, runs a shell command, or receives a tool response, Nightfall scans the content and can block the action.

* Supported agents: Claude Code, Cursor, VS Code Copilot
* Enforcement: Block or Monitor on prompts, tool calls, and shell commands. Tool responses can also be blocked on the macOS and Windows agents (replacing the tool output before the model reads it; not supported on VS Code Copilot). Model responses are monitor only

#### OpenTelemetry, async monitoring

OpenTelemetry captures agent activity after actions complete: session audit, cost, model, and tool activity. It does not block.

* Supported agents: Claude Cowork
* Additional data: token usage, cost per prompt, model name, API errors

#### Claude Inference Hooks

For Claude Enterprise, Nightfall can allow or deny a prompt before the model runs. This covers claude.ai, Cowork, and Claude Code signed into that organization, including machines with no agent. It does not replace device hooks when the shell command or the tool call itself must not start. Setup and policy templates are on [Hook policy recommendations](/data-exfiltration-prevention/ai-agent-security/ai-governance/hook-policy-recommendations.md).

### Prerequisites

Device hooks need the following. Inference Hooks do not.

#### Nightfall endpoint agent

* Version 1.2.12.11 or later.
* The agent must be installed and running on each endpoint where AI agents are used.

#### MDM configuration profile

{% hint style="info" %}
Deploy the Nightfall endpoint agent configuration profile (v3 or later) via your MDM provider. This profile grants the agent its OS permissions. It is not the Claude Code hooks file `nightfall-hooks.mobileconfig`.
{% endhint %}

#### AI agent policy

At least one AI Agent Security policy must be active. Install hooks from the client guides under Setup and Installation. Claude Code, Cursor, and VS Code Copilot each have their own package and path.

### Next steps

* [Choosing a control](/data-exfiltration-prevention/ai-agent-security/ai-governance/choosing-a-control.md)
* Setup and Installation
* Hooks vs. Open Telemetry
* [Hook policy recommendations](/data-exfiltration-prevention/ai-agent-security/ai-governance/hook-policy-recommendations.md)


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://help.nightfall.ai/data-exfiltration-prevention/ai-agent-security/ai-governance/auditability-and-control.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.