Knostic
Knostic secures agentic workflows and AI-native development so enterprises can accelerate AI adoption safely. Its platform discovers and secures AI agents and coding assistants, controls data exposure and unsafe actions, and assesses the associated agentic supply chain, including MCP servers, skills, IDE extensions, and rules.
At a Glance
- Enterprise security teams and CISOs
- Engineering and platform teams deploying AI coding assistants
- Software and SaaS providers
- Financial services and banks
- +2 more
AI Tools by Knostic
(1)
Knostic
AI Coding Agent Security Platform
Discussions
No discussions yet
Be the first to start a discussion about Knostic
Latest News
Auto-patch for OpenAnt: Generating trust in autonomous patching
The Duck Song: Malicious Solidity Extensions Hide a Cross-Platform Payload in an MP4
When ‘Auto-Signing’ Sends Your Wallet: a malicious MCP package on npm
SaassyCode Repackaged: Four New VS Code Extensions and a New Delivery Chain
Products & Services
Security for AI coding agents and IDEs. Kirin runs at the agentic endpoint, validates connections, inspects prompts, tool calls, dependencies, extensions and MCP connections, enforces policy inline, detects secrets and prompt injection, and blocks unsafe or destructive behavior in real time.
Agentic supply-chain threat intelligence and scanning for AI agent skills, MCP servers, and VS Code extensions. It discovers and tracks components, scans for prompt injection and supply-chain threats, provides dangerous/risky/clean verdicts, supports SHA-256 lookup, and exposes a REST API.
Free, open-source LLM-based vulnerability discovery for proactively finding verified security flaws while minimizing false positives and false negatives. It uses code parsing, call-graph and reachability analysis, agentic exposure classification, vulnerability discovery, constrained exploitability verification, and dynamic sandboxed testing. A managed offering is available via waitlist.
Organization-wide visibility into sanctioned and unsanctioned AI tools, agents, local LLMs and AI services, with risk scoring and team mapping.
Market Position
Knostic positions itself as a control plane for agentic AI on the endpoint, combining agent-layer enforcement with shadow-AI discovery and agentic supply-chain intelligence. It differentiates from conventional EDR and network monitoring by inspecting prompts, tool calls, configurations and MCP traffic inside applications; relevant competitive alternatives include Palo Alto Networks Cortex/Koi and other AI security, AI-SPM, agent security and MCP/IDE supply-chain security vendors. Knostic's OpenAnt explicitly says it does not intend to compete with OpenAI Codex Security or Anthropic Claude Code Security.
Leadership
Founders
Gadi Evron
Cybersecurity expert and serial entrepreneur; founded the ACoD cybersecurity conference, founded Cymmetria and led it as CEO until its acquisition, led PwC's Cyber Security Center, and served as CISO for Israel's National Digital Authority.
Sounil Yu
Cybersecurity leader and creator of the Cyber Defense Matrix and DIE Triad; previously Bank of America's Chief Security Scientist and CISO at JupiterOne. He is also a CMMC Cyber Accreditation Body board member, Global Resilience Federation board member, GMU National Security Institute senior fellow, and Cybersecurity Hall of Fame inductee.
Executive Team
Gadi Evron
Co-Founder and CEO
Cybersecurity expert and serial entrepreneur; previously founded Cymmetria, led PwC's Cyber Security Center, founded ACoD, and served as CISO for Israel's National Digital Authority.
Sounil Yu
Co-Founder and CTO
Creator of the Cyber Defense Matrix and DIE Triad; previously Bank of America's Chief Security Scientist and CISO at JupiterOne.
Founding Story
Knostic was started in 2023 by Gadi Evron and Sounil Yu to make enterprise AI adoption safe. The initial insight was that LLMs and AI assistants lacked the contextual ability to respect need-to-know boundaries and could overshare sensitive information; the company set out to provide controls that let enterprises use AI without compromising security.
Business Model
Revenue Model
Recurring per-user software subscriptions for Kirin and AgentMesh, with custom enterprise pricing; AgentMesh also offers REST API access and OpenAnt is free open source with a managed offering planned/available via waitlist.
Pricing Tiers
For individual developers and teams of 1–5 users; continuous project scanning, granular policies, configuration enforcement, extension/rules/generated-code/MCP scanning and standard support.
Adds unified dashboard, organizational policy, RBAC, audit logs, cross-organization MCP/rules/hooks visibility and priority support; advanced contextual detection is an add-on.
Adds policy groups, advanced contextual detection, enterprise SSO, SIEM and Slack integrations, compliance-ready reporting and dedicated support.
Login required; 300 items per category, web UI and limited API access.
Full catalog, up to 300 API calls per month (20/day), with Kirin IDE security included.
Full catalog and unlimited API access, with Kirin IDE security included.
Target Markets
- Enterprise security teams and CISOs
- Engineering and platform teams deploying AI coding assistants
- Software and SaaS providers
- Financial services and banks
- Healthcare, energy and other regulated industries
- Private-equity portfolio companies and organizations scaling AI adoption
- Secure enterprise adoption of AI coding agents and copilots
- Prevent data exfiltration, secrets leakage and LLM oversharing
- Govern shadow AI and inventory AI tools used by employees
- Stop destructive agent commands and risky tool calls before execution
- Vet MCP servers, skills, extensions and packages for supply-chain threats
- Find vulnerabilities in AI-generated or open-source code