Straiker

Straiker: security for the AI agents your enterprise is deploying

Straiker is an agentic AI security company that treats AI agents as a new workforce that needs its own security stack. The platform delivers three integrated capabilities: Discover AI maps the agents, MCP servers, and AI workflows running across an enterprise and monitors their posture; Ascend AI is an adversarial red-teaming engine that attacks those agents before deployment to expose prompt injection, goal hijacking, tool misuse, and inter-agent manipulation; and Defend AI provides runtime protection that detects and blocks identity abuse, memory poisoning, data exfiltration, and resource exploitation as agents operate. A dedicated Secure Claude Code offering covers coding-agent environments.

It is built for Fortune 500 enterprises and frontier AI labs rolling out coding agents, custom-built agents, and productivity agents in industries such as financial services, healthcare, technology, media, and retail. The stakes are concrete: in testing by STAR Labs, Straiker's threat research arm, 36% of successful attacks on coding agents achieved remote code execution and 91% of attacks on productivity agents led to silent data exfiltration.

Key Features

• Discover AI: inventory and posture monitoring for agents, MCP servers, and AI workflows, including misconfiguration detection
• Ascend AI: continuous adversarial red-teaming powered by proprietary threat data
• Defend AI: runtime detection and blocking without slowing agent performance
• Secure Claude Code for coding-agent environments
• STAR Labs threat research feeding new attack techniques back into the product
• SOC 2 and ISO 27001 certified; OWASP AI Exchange gold sponsor

Ideal Use Case

Straiker fits security teams at large enterprises that have moved past AI pilots and now run agents with real access to tools, data, and credentials. If your organization is deploying coding agents or customer-facing agents and your existing AppSec stack cannot see what they are doing, this is the category Straiker was built for.

How Straiker differentiates

Straiker covers the full lifecycle — discovery, pre-deployment testing, and runtime defense — rather than a single guardrail layer. It announced a $64M Series A in June 2026 led by Marathon Management Partners with Citi Ventures, Illuminate Financial, and Workday Ventures, plus continued backing from Bain Capital Ventures and Lightspeed, bringing total funding to $85M. Customers include Comcast, DirectTV, Fortinet, Snowflake, Deloitte, Coupang, and American Express Global Business Travel. CEO Ankur Shah previously ran Palo Alto Networks' Prisma Cloud business; CTO Sreenath Kurupati founded fraud-detection company Cyberfend, acquired by Akamai.

FAQ

What is Straiker? An agentic AI security platform that discovers enterprise AI agents, red-teams them before deployment, and protects them at runtime.

What threats does Straiker address? Prompt injection, goal hijacking, tool misuse, memory poisoning, identity abuse, data exfiltration, and inter-agent manipulation.

How much does Straiker cost? Pricing is enterprise sales-led with no published tiers; contact Straiker for a quote.

Who is behind Straiker? Founders Ankur Shah (ex-Palo Alto Networks Prisma Cloud SVP/GM) and Sreenath Kurupati (Cyberfend founder, ex-Akamai), backed by Marathon, Bain Capital Ventures, and Lightspeed.

tl;dr

Straiker secures enterprise AI agents end to end: discovery, adversarial red-teaming, and runtime blocking. $64M Series A in June 2026 with customers including Comcast, Fortinet, and Snowflake.

Related

Looking for more options? Browse the Security & Governance directory or read our best AI security tools listicle. Straiker is also tracked on Crunchbase.