Tools
MessageGuard ML
By OpenClaw plugin: ML-based secret detection using DistilBERT/ONNX via transformers.js
Install
npm install @andrewlabs/openclaw-messageguard-mlREADME
# @andrewlabs/openclaw-messageguard-ml
ML-powered companion plugin to `@andrewlabs/openclaw-messageguard`.
This plugin uses a DistilBERT token classification model via `@huggingface/transformers` (ONNX Runtime under the hood) to detect sensitive content in outgoing messages and replace detected spans with `[REDACTED]`.
## Features
- OpenClaw `message_sending` hook integration
- Automatic model download from Hugging Face on first use (then cached locally)
- Fails open: if model download/inference fails, message passes through unchanged
- Configurable model id, confidence threshold, and redaction token
## Install
```bash
npm install @andrewlabs/openclaw-messageguard-ml
```
Ensure your OpenClaw plugin loader can discover package extensions via `openclaw.extensions`.
## Manifest
The package includes `openclaw.plugin.json` with plugin id `messageguard-ml` and configuration schema.
## Configuration
`openclaw.plugin.json` supports:
- `enabled` (boolean, default `true`)
- `modelId` (string, default `AndrewAndrewsen/distilbert-secret-masker`)
- `threshold` (number, 0-1, default `0.5`)
- `mask` (string, default `[REDACTED]`)
## How it Works
1. On plugin startup, `message_sending` hook is registered.
2. For each outgoing message, the model runs token classification.
3. Spans predicted as sensitive at/above threshold are merged and redacted.
4. Sanitized content is returned to OpenClaw.
If model loading or inference fails (for example, model repo not yet available), the plugin logs a warning and returns without modifying the message.
## Exporting/Quantizing ONNX
Use the helper script:
```bash
python3 scripts/export_onnx.py
```
Optional push to Hugging Face (requires token):
```bash
HF_TOKEN=... python3 scripts/export_onnx.py --push
```
If `AndrewAndrewsen/distilbert-secret-masker` does not exist on Hugging Face, the script exits with a clear warning.
## Comparison With Regex Plugin
- `@andrewlabs/openclaw-messageguard`: deterministic regex rules and policy actions
- `@andrewlabs/openclaw-messageguard-ml`: learned token classification for broader, context-sensitive detection
Running both can provide layered defense.
## Development
```bash
npm install
npm run build
```
tools
Comments
Sign in to leave a comment