Search

Deterministic, no-LLM SecurityAnalyzer for OpenHands, backed by the rein engine.

Code review agent powered by DSPy

Public Preview — GitHub Copilot Extension for agent governance code review: detects missing policy checks, unguarded tool calls, and audit logging gaps

Policy-aware AI permission reviewer for OpenCode. Evaluates tool and command requests, then allows once, denies with feedback, or escalates to the user.

Vibe Hacker — Terminal AI cybersecurity assistant. Free models, autonomous agent, multi-provider rotation.

Arc Gate runtime governance for CrewAI agents

Open source computer-use agent — control your computer with natural language or voice

Score (query, document) pairs for vector/RAG poisoning signals: vector-text mismatch, instruction-like payloads, NaN, suspiciously round numbers. Python port of @mukundakatta/vector-poison-score.

ColoMind Charter System - Define AI capability boundaries with document bindings

Wrap your Anthropic / OpenAI Python client. Every tool call is inspected by Clavenar before it runs.

AI model: AI45Research/AgentDoG1.5-Qwen3.5-2B (text-classification)

Prompt injection firewall for LLM agents. Scan tool outputs and user messages before they reach your model.

Ask your AI coding agent, in natural language, to search production logs or inspect a database — safely. Two drop-in MCP servers (MySQL, SSH logs) keep safety red lines in code: read-only by default, dangerous SQL/shell blocked, production disabled, no stored credentials. A recommended Skill layer teaches the model when and how to use them.

AI model: mradermacher/AgentDoG-FG-Qwen2.5-7B-GGUF

面向 AI Agent 的授权网络评估工具集 — TCP 端口扫描 / DNS 子域枚举 / 目录爆破 / HTTP 指纹识别，JSON in / JSON out 并附 MCP 服务（BIT 生态）

Public Preview — Episodic Memory Kernel for AI agent experience storage

Cloudflare OS is Cloudflare’s open source AI productivity environment for building company-specific agent workspaces. It combines an agent chat UI, generated workspace apps called gadgets, Workers-based sandboxing, and Gatekeepers that mediate access to external systems with narrow permissions, logging, and human approval.

Policy-aware AI permission reviewer for OpenCode V1 and V2: auto-approves routine agent tool calls, denies risky ones with feedback, and escalates the rest to you

Safe execution layer for OpenClaw/Clawbot using tjs-lang's capability-based VM

Catches AI hallucinated npm packages before an agent installs them