# AI agents should not have risky HTTP configurations

AI agents should not have risky HTTP configurations. A Maester check for Copilot Studio, Agents.

Source: https://maester.dev/docs/tests/MT.1115

<!-- This file is generated by website/scripts/generate-test-docs.mjs. Do not edit manually. -->

<div class="test-byline"><div class="test-byline-avatars"><a class="test-byline-avatar test-byline-avatar--author" href="/contributors/lnfernux" title="Truls Thorstad Dahlsveen · Original author"><img src="https://github.com/lnfernux.png" alt="Truls Thorstad Dahlsveen" width="28" height="28" loading="lazy" decoding="async" /></a></div><div class="test-byline-meta"><span class="test-byline-text">Contributed by <a href="/contributors/lnfernux">Truls Thorstad Dahlsveen</a></span><a class="test-byline-link" href="/contributors">All contributors →</a></div></div>

<p class="test-part">Part of <a href="/docs/tests/maester">Maester</a> · Copilot Studio · Agents · <a href="/docs/tests">All checks</a></p>

```powershell title="Run this check"
Invoke-Maester -TestId MT.1115
```

## Overview

AI agents should not use risky HTTP configurations.

Agents with HTTP request nodes in topics connecting to non-standard ports or using plain HTTP (instead of HTTPS) may be misconfigured or could indicate data exfiltration or command-and-control communication channels.

### How to fix

Review the HTTP request nodes in each flagged agent's topics. Ensure all HTTP requests use HTTPS on standard port 443. Replace direct HTTP calls with Power Platform connectors where possible, as connectors provide built-in governance and DLP policy enforcement.

Learn more: [Configure data policies for agents](https://learn.microsoft.com/microsoft-copilot-studio/admin-data-loss-prevention?tabs=webapp#block-http-requests)

## Details

<details>
<summary>Every attribute of this check</summary>

| Field | Value |
| --- | --- |
| Test ID | MT.1115 |
| Severity | Medium |
| Suite | Maester |
| Product | Copilot Studio |
| Feature | Agents |
| Area | AI and agents |
| Resource type | Agent |
| User impact | Low |
| Implementation cost | Medium |
| MITRE ATT&CK | T1571 Non-Standard Port, T1048 Exfiltration Over Alternative Protocol, T1071.001 Web Protocols |
| Supports controls | nist-800-53-r5:SC-7 |
| PowerShell test | [Test-MtAIAgentRiskyHttpConfig](https://github.com/maester365/maester/blob/main/tests/maester/copilot-studio/Test.MT.1115.ps1) |
| Services | Graph |

</details>

## Source

- Test: [`tests/maester/copilot-studio/Test.MT.1115.ps1`](https://github.com/maester365/maester/blob/main/tests/maester/copilot-studio/Test.MT.1115.ps1)
- Documentation: [`tests/maester/copilot-studio/Test.MT.1115.md`](https://github.com/maester365/maester/blob/main/tests/maester/copilot-studio/Test.MT.1115.md)
