Optional ONNX ML classifier for ai-shield — DeBERTa-style prompt-injection detection alongside heuristic patterns

ML classifier add-on for Agent Armor. Downloads and runs a DeBERTa-v3-small ONNX model for agent trap detection.

Privacy-first OpenClaw plugin. Classifies and routes every message by sensitivity, blocks prompt injections locally, and keeps your AI spend in check. Built by Centrase AI.

Run open-jev- and Laya-shaped typed-decision models in the browser: one state plus typed questions in, calibrated probabilities out.

Pro plugin pack for Bastion AI Gateway

3-layer prompt injection defence for chat bots

Detect prompt injection attacks using the qualifire/prompt-injection-sentinel model

OpenCode plugin that scans tool outputs using Meta's Llama Prompt Guard 2

Sentiment and regime overlay for trading bots responding to policy-driven shocks.

Focused MCP server for persistent memory capabilities - semantic search, knowledge graphs, and session context

Local cross-encoder reranker for Engram — runs mxbai-rerank-v1 via ONNX Runtime (no API calls)

Workers AI email classifier adapter for @rafters/mail

TCL NLI scoring service (Hugging Face style HTTP API).

Local prompt injection and jailbreak detection for LLM applications

Visualize BMAD Method epics, stories & documents with AI-powered git commit matching

TypeScript SDK for MachinaFi — Autonomous AI Agents on Base L2

A JupyterLab extension for academic research management and WWC quality assessment

Cerberus — a local-first security gateway for autonomous AI coding agents. Intercept, risk-score, and human-approve every tool call. Terminal-first, with a forensic dashboard.

Typed decisions in the browser: run the open-jev model (a Jev-shaped System One model) with Transformers.js.

Local ONNX decision models for OpenClaw