[ECCV 2024] The official code for "AdaShield: Safeguarding Multimodal Large Language Models from Structure-based Attack via Adaptive Shield Prompting."

Shields your confidential data in LLM prompts from third party AI providers, allowing you to send with confidence without compromising security and privacy.

Self-learning prompt injection detection engine — 25 input detectors (10 languages), 5 output scanners, PII redaction, red team self-testing, F1: 96.0% with 0% false positives. Docker, GitHub Action, pre-commit, FastAPI/Flask/Django/LangChain/CrewAI/Dify/n8n.

prompt quality management for Claude Code

OpenClaw Prompt Shield: Security framework that protects OpenClaw AI agents from prompt injections, data leaks, and dangerous commands using Azure AI Content Safety and Microsoft Purview DLP.

Fast prompt injection detection engine. Rust library + CLI + WASM.

Real-time prompt injection, jailbreak and exfiltration detection middleware for LLM APIs. Zero ML dependencies. Sub-2ms latency. 75 tests.

No description available

Security middleware that shields AI applications from prompt injection, jailbreaking, and obfuscated attacks through composable defense layers.

The full-stack safety layer for AI agents. Budget limits, prompt injection shields, PII filtering, output firewalls, and hooks — in 2 lines of code.

Open-source semantic security shield that inspects, sanitizes, and blocks malicious prompts to prevent jailbreaks, data exfiltration, and unauthorized tool use in conversational agents.

Browser extension that detects PII in AI chat inputs and lets you redact before sending.

A 4-class classification system for securing technical LLMs

🛡️ Protect your LLM apps from prompt injection attacks. 3 lines to integrate, Detects jailbreaks, prompt extraction, instruction overrides & more.

Lightweight prompt injection detection for Python. Fast, easy integration with LangChain and AutoGen.

Production-ready security playbook for AI agent deployments. Addresses backup file persistence, authentication bypass, and prompt injection vulnerabilities. Integrates openclaw-detect, openclaw-telemetry, openclaw-shield, and clawguard.

No description available

No description available

This repository contains a simple but functional Chrome Extension. It monitors input fields, textareas, and editable elements on websites. When forbidden words are detected, the extension warns the user (or blocks submission). Users can manage the blocked words list from the popup or options page.

No description available

No description available

> PromptShield is a prompt security middleware for LLMs. It uses a fine-tuned DistilBERT classifier to label incoming prompts as safe, unsafe, suspicious, or jailbreak before forwarding to Google Gemini 2.5 Flash. Built with Python, Transformers, and a vanilla HTML/CSS/JS frontend.

Runtime security firewall for LLM applications.

A Context-Aware Prompt Injection Defense System for LLM Chatbots that detects and neutralizes prompt injection attacks before they reach your LLM.

No description available

No description available

Lightweight prompt injection detector. 22 attack patterns. Blocks jailbreaks before they reach your model.

Security audit skill for Claude Code — 18 automated checks for prompt injection, data exfiltration, supply chain attacks, and persistence mechanisms in skills, MCP servers, and npm packages.

making a project to filter prompts to prevent from prompt injection

No description available