# Knostic

> Security platform that discovers, monitors and enforces policy on AI agents, coding assistants and their supply chain at the endpoint.

Knostic builds security software for agentic AI workflows, covering coding assistants such as Cursor, Claude Code, GitHub Copilot and Windsurf, as well as Claude Cowork. Its products are Kirin, an endpoint security layer for agents, and AgentMesh, threat intelligence for skills, MCP servers and extensions. Knostic also publishes the open-source OpenAnt vulnerability scanner.

## What It Is

Knostic is an agentic-endpoint security platform. Instead of asking a model to behave, it checks each agent action against a central policy before the action runs. Allowed actions proceed, while risky ones are blocked or routed to a human for review. The same policy applies across the different agents a team uses.

## How Kirin Works

Kirin installs as a user-space client (no admin rights required) and, for Cursor and VS Code, as an extension. Policies run in Off, Monitor or Enforce modes. In Enforce mode, Kirin can block a read of secret files or sanitize secrets before they reach the model. It quarantines unverified MCP servers for admin review and enforces allow/block lists for extensions. Kirin also provides shadow AI discovery of local LLMs, browser chatbots and unsanctioned APIs, along with centralized policy management. According to the FAQ, supported agents are Claude Code CLI, Claude Desktop and Cowork, Cursor, Windsurf, and VS Code with GitHub Copilot, on macOS and Windows, with Linux in beta.

## AgentMesh and the Agentic Supply Chain

AgentMesh continuously discovers and scans agent skills, MCP servers and VS Code extensions. It returns dangerous, risky or clean verdicts, supports lookup by SHA-256, and offers a REST API. Knostic's research blog reports findings on malicious VS Code extensions and MCP packages.

## AI SDLC and Generated Code

Knostic describes detecting hallucinated and vulnerable packages, enforcing secure coding rules including automatic threat modeling, and discovering vulnerabilities and generating patches. The FAQ says audit logs and SIEM, Slack and Teams integrations are available on higher tiers, with Enterprise SSO.

## Features
- Shadow AI discovery
- Inline policy enforcement for agent actions
- Secrets leakage blocking
- MCP server quarantine and review
- Extension allow/block lists
- Centralized security posture management
- Agentic supply chain scanning
- SHA-256 artifact lookup
- Hallucinated and vulnerable package detection
- Automatic threat modeling for AI-generated code
- Audit logs and SIEM integration

## Integrations
Cursor, Claude Code, Claude Cowork, GitHub Copilot, Windsurf, VS Code, Slack, Microsoft Teams, SIEM, MDM

## Platforms
WINDOWS, MACOS, LINUX, WEB, API, VSC_EXTENSION, JETBRAINS_PLUGIN, CLI

## Pricing
Freemium — Free tier available with paid upgrades

## Links
- Website: https://www.knostic.ai
- EveryDev.ai: https://www.everydev.ai/tools/knostic