Open detection standard -- like Sigma, but for AI agents. Executable rules for prompt injection, tool poisoning, context exfiltration, and MCP attacks. Shipped in Cisco AI Defense. MIT-licensed.

Thin wrapper around agent-threat-rules — re-exports all detection logic for monorepo consumers.

OpenGuardrails (OGR) conformant detector backed by Agent Threat Rules (ATR) — accepts a GuardEvent, returns a Verdict, using the open ATR rule engine.

Security audit plugin for Claude Code — scans skills, hooks, and MCP configs for prompt injection, data exfiltration, supply chain attacks, and 37 threat patterns. v3.1.1 bundles Agent Threat Rules (ATR) v2.1.2 as a reference resource.

Cross-realm type predicates for Bare

OCI NodeJS client for Threat Intelligence Service

Sandbox-enforcing implementation of the DeepSeek Harness filesystem seam: fences write/edit by the per-call sandbox mode (read-only denies mutation, workspace-write contains it to the workspace + temp roots) while reads pass through

Heuristic lexer for detecting imports and exports in JavaScript modules

Panguard AI - Unified security CLI

Structured cloning algorithm for JavaScript

Write-time threat guard for evolution tools (community build)

System logger for Bare

Your agent writes bad React. This catches it

Ship compositional knowledge for AI coding agents alongside your npm packages

MCP Proxy — runtime interception for AI agent tool calls using ATR rules

Genx CLI

Find geolocation data from IP addresses (e.g. city, country, timezone) using the IPLocate.io API

TypeSpec Azure Resource Manager library

Coding agent CLI with read, bash, edit, write tools and session management

A programmable agent-first linter for Tailwind design systems.