> For the complete documentation index, see [llms.txt](https://ai-security-docs.akto.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://ai-security-docs.akto.io/akto-argus-agentic-ai-security-for-homegrown-ai/prompt-hardening.md).

# Prompt Hardening

## Overview

Prompt Hardening supports AI Red Teaming of AI agents against prompt injection and related adversarial techniques. Prompt Hardening focuses on identifying unsafe responses, policy violations, and unintended agent behaviour triggered by malicious prompts. This capability helps enterprise security and platform teams validate AI agent resilience before production exposure.

Prompt Hardening is part of the **Agentic Security** product and supports controlled, repeatable testing workflows for AI-driven systems.

<figure><img src="/files/dvQIGowMWQTLW4tTIdK7" alt="" width="563"><figcaption></figcaption></figure>

## Prompt Types Available for Scanning

Prompt Hardening provides structured prompt sources to support both targeted and exploratory scanning.

### Custom Security Prompts

Custom Security Prompts allow you to probe specific attack classes with precise intent.

1. **Tool Misuse**\
   Evaluates whether an AI agent invokes tools or actions beyond intended authorisation.
2. **Data Leakage**\
   Probes whether an AI agent exposes sensitive, private, or restricted data.
3. **Prompt Injection**\
   Assesses whether an AI agent follows malicious instructions embedded in user input.
4. **Policy Evasion**\
   Validates whether an AI agent bypasses internal governance, safety, or compliance controls.
5. **Jailbreak & Safety Bypass**\
   Probes attempts to override system-level safeguards and safety boundaries.

### Akto Default

The Akto Security Prompt Library provides curated prompts aligned with common enterprise AI risk scenarios.

* **Security Probing**\
  Focuses on vulnerability discovery and unsafe agent behaviour.
* **Performance Testing**\
  Evaluates response stability, consistency, and behaviour under edge conditions.

## Next Step

After reviewing available prompt types, you can proceed to [hands-on scanning and configuration](/akto-argus-agentic-ai-security-for-homegrown-ai/prompt-hardening/play-around-in-prompt-hardening.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 dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://ai-security-docs.akto.io/akto-argus-agentic-ai-security-for-homegrown-ai/prompt-hardening.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
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.