SecureVector

Hermes Agent, Guarded. Every tool call audited, threats blocked in enforce mode, 79 built-in tools under your policy. Nothing leaves your machine.

AI regulation is coming - the EU AI Act leads, others will follow. You can't govern agents you can't see. SecureVector v4.8.0: device-level Agent Governance posture, free & open source. Fleet-wide posture with a score in the cloud app. Posture, not a compliance badge. https://www.youtube.com/watch?v=G8-JQ-nRmTY

- Guard plugin and hooks for Cursor and Guthub Copilot CLI. - Cloud app: consolidated live agent map + real-time threat detection, fleet-wide in the cloud app. - SV Guardian ML model : A lightweight ML model powers SecureVector, also on PyPI. v1.3.0 sees through base64/hex/URL encoding and long-document burying. Trained 100% on our own rules corpus. pip install securevector-guardian-model. Already packaged with securevector-app checkout GitHub repo https://github.com/Secure-Vector/securevector-ai-threat-monitor Demo: https://youtu.be/z2M4jSqP4Aw?si=3DVjLnaRxjFl9Y7j

SecureVector Guard installs as a native plugin into Claude Code, Codex, and OpenClaw — and gives you one place to see and control what your agents actually do: tool-permission enforcement, tamper-evident audit, prompt-injection detection, a live Agent Map, and per-run Agent Runs traces. Check Demo: https://youtu.be/9RByIHSV95s?si=fRX5ExVUX0GPy2oE

→ Bidirectional tool-response scans - IDPI + PII rules on MCP/WebFetch/Read/Grep responses → MCP Tool Inventory (Bill of Tools) - SBOM-style supply-chain view along with export options. → Secret Detections - hash-only audit log → PEM private-key + OpenSSH-binary detection (T1552.004)

Launched SecureVector v4.2.1 - SecureVector plugin for Claude Code — every MCP tool call in your statusline, audited locally. - MCP Catalogue + Cloud Policies — author policies to allow/deny/audit tools, push it across every enrolled developer machine (or multiple devices) on next sync. MCP policy + Claude code plugin Demo: https://www.youtube.com/watch?v=V25yxqlAoY0 Device Enrollment Demo: https://youtu.be/vHXWl38hRjM

New in SecureVector: cloud-managed MCP policies and tool permissions, synced across every device where your agents run. Push one policy from the dashboard and it takes effect everywhere — via the native plugin for OpenClaw (the only plugin available today, Claude Code plugin coming next) or via the multi-provider proxy for, n8n, LangChain, CrewAI, and any OpenAI-compatible setup. Manage a fleet of dev machines without touching each one.

Official n8n node now supports both local and cloud mode. https://n8n.io/integrations/securevector/

Founders, Indie Devs, SOCs, CISOs — check this out! Forward every AI agent threat + tool audit to Splunk, Datadog, Sentinel, Grafana, or any webhook. AI Agents → SecureVector → Your SIEM → Done Works with LangChain, OpenClaw, Ollama, n8n. Would love feedback.

SecureVector native plugin for OpenClaw. Give it a try! Using OpenClaw integration? no proxy needed. Install Native plugin and start observing and auditing. No latency.

SecureVector v3.2.0 just shipped! Introducing Skill Scanner - scan any AI skill before you install or run it. Your agent supply chain deserves better security. https://youtu.be/KWN6uxNX3sw?si=ez-UYVLCu2jwWTSk

SecureVector v3.0.1 — a local, real-time, open-source AI firewall. I've been working on this for the past couple of weeks. The goal is simple: give the community a free security layer and budget tracking for OpenClaw/AI agents. GitHub: [github.com/Secure-Vector/securevector-ai-threat-monitor](http://github.com/Secure-Vector/securevector-ai-threat-monitor) Demo: [https://www.youtube.com/watch?v=tLVDvHIm-0c](https://www.youtube.com/watch?v=tLVDvHIm-0c) **What's new in v3.0.1:** * Agent Tool Permissions — allow/block agent tool calls * Cost Tracking & Budget Limits — per-agent spend + global daily budget * 28 new threat detection rules (77 total)

OpenSource AI firewall, check out https://github.com/Secure-Vector/securevector-ai-threat-monitor.

Revamp website with new chat feature for detection of threats TEST THREATS INSTANTLY WITH CHAT Test any prompt or file upload for security threats in real-time. Our interactive chat interface analyzes inputs for prompt injections, jailbreaks, and data exfiltration attempts. See instant risk scores, confidence levels, and detailed threat breakdowns.

We’ve shipped several powerful new features to strengthen your AI security workflows: • New n8n Node Support SecureVector now integrates directly with n8n, making it easy to protect agentic workflows and automations with built-in AI security scanning. • Alerts & Notifications Stay informed in real time. Enable alerts and get notified instantly when potential attacks or risky behaviors are detected in your n8n workflows or AI applications—so you can act before issues escalate. • Custom AI-Generated Rules Create industry-specific security rules on demand. Our AI-powered rule generator helps you define tailored protections that match your use case, without manual rule writing.