Search

Guard LangChain.js tools with EMILIA Protocol — require an offline-verifiable authorization receipt (EP-RECEIPT-v1) before an irreversible tool runs: missing -> refused, valid -> runs, replay -> refused, forged -> refused (RR-1). The approval becomes port

Responsible AI auditing for LLMs and SLMs — deception, fairness, sociotechnical risk, regulatory compliance

Audit Agent Skills and packages before they touch your machine. Deterministic checks. Exit codes. No vibes.

SecureVector SDK for CrewAI — brings the local threat monitor's three controls (tool-call permissions, secret/data-leak detection, threat detection) to every CrewAI tool call, with tamper-evident audit logging.

Node SDK for checking AI agent tool calls with Grimdall.

Bound what a compromised LLM agent can reach and do

The Open Authorization Standard for AI Agents. Framework-agnostic tool permissions, identity verification, scoped access control, and audit logging for any AI agent.

OWASP Foundation web repository

Runtime risk scoring for AI agent tool calls. Detect privilege escalation, baseline deviations, and data exfiltration inline — under 1ms, zero dependencies, no LLM.

Defensive overlay for AI agents — 8-layer architecture with harness-agnostic adapters; ships 6 of 8 layers (1 skill_vetting, 2 sanitize, 3 structured_output, 4 bash_guard+write_guard, 6 audit, 7 config) plus adapters for Claude Code (PreToolUse) and OpenClaw (before_tool_call)

The Pydantic-based Firewall for MCP Servers. Stops hallucinated tool calls, validates schemas, and sandboxes dangerous operations.

Defensive LLM agent honeypot & token tarpit - FastAPI backend with an OpenAI-compatible endpoint that wastes hostile agents' context windows.

AI Agent 红队安全测试平台 — 从 CLI 到 Dashboard

A portable coding lifecycle and autonomous repository steward for AI agents with deterministic routing across 44 skills, durable state, mutation-aware verification, continuous CI healing, specialized skill packs, lifecycle hooks, DeepSeek Harness plugin &

Official Python SDK for AgentGuard Intelligence Engine and Action Firewall.

Static analyzer and GitHub Action for detecting Agentic Workflow Injection risks in AI-powered GitHub Actions workflows.

SecureVector SDK for LangChain — brings the local threat monitor's three controls (tool-call permissions, secret/data-leak detection, threat detection) to every LangChain tool call, with tamper-evident audit logging.

Authensor guardrail adapter for LangChain/LangGraph

Turn the OpenAI Agents SDK human-in-the-loop tool-approval step into a verifiable, offline-checkable EMILIA authorization receipt. OpenAI pauses an agent and asks for approval; EMILIA makes that approval portable, tamper-evident evidence — a named human a

LLM Prompt Injection Detection CLI — 3-layer detection (Vigil + DistilBERT ONNX + Rules)