> For the complete documentation index, see [llms.txt](https://ai.ortusbooks.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://ai.ortusbooks.com/main-components/middleware/input-sanitizer.md).

# InputSanitizerMiddleware

Heuristic prompt-injection scanning and unicode hygiene for inbound content and tool results

Class: `bxModules.bxai.models.middleware.security.InputSanitizerMiddleware`

Scans inbound user content and optional tool results for prompt-injection indicators. Applies unicode normalization and invisible-character stripping before model execution.

## Features

* Scans user-role content before LLM call
* Optional scan of tool and MCP results in `wrapToolCall`
* Unicode hygiene via normalization and zero-width stripping
* Actions: `block`, `strip`, `flag`, `log`
* Findings can be stamped to `chatRequest.providerOptions.securityFindings`

## Constructor

```javascript
middleware = new bxModules.bxai.models.middleware.security.InputSanitizerMiddleware(
    action          : "flag",
    detectors       : [],
    customPatterns  : [],
    normalizeUnicode: true,
    stripZeroWidth  : true,
    scanToolResults : true
)
```

## Configuration

| Option             | Type    | Default  | Description                                      |
| ------------------ | ------- | -------- | ------------------------------------------------ |
| `action`           | string  | `"flag"` | Handling mode: `block`, `strip`, `flag`, `log`   |
| `detectors`        | array   | `[]`     | Detector names to run; empty means all built-ins |
| `customPatterns`   | array   | `[]`     | Additional patterns: `[ { name, regex } ]`       |
| `normalizeUnicode` | boolean | `true`   | Apply NFKC normalization                         |
| `stripZeroWidth`   | boolean | `true`   | Remove zero-width and invisible characters       |
| `scanToolResults`  | boolean | `true`   | Scan tool result strings for indirect injection  |

## Hooks Used

* `beforeLLMCall`
* `wrapToolCall`

## Action Semantics

* `block`: throws `BXAI.SecurityViolation` on findings before LLM call; tool-result findings return blocked marker text
* `strip`: removes matched fragments and continues
* `flag`: keeps content, stamps findings in provider options, logs warnings
* `log`: logs warnings only

## Example

```javascript
agent = aiAgent(
    middleware: [
        new bxModules.bxai.models.middleware.security.InputSanitizerMiddleware(
            action        : "block",
            detectors     : [ "instructionOverride", "jailbreak" ],
            customPatterns: [ { name: "internalCodes", regex: "(?i)PROJ-\\d{4}-SECRET" } ]
        )
    ]
)
```

## Notes

* Invalid actions throw `InvalidAction` during construction.
* Message content is mutated in place, including multipart text content sections.


---

# 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://ai.ortusbooks.com/main-components/middleware/input-sanitizer.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.