Agents have made the entire enterprise operating system the attack surface. Straiker gives security teams the control plane to deploy agents securely.
SAN FRANCISCO, March 23, 2026 /PRNewswire/ -- RSAC 2026 – Straiker, the fastest-growing agentic security company, today launched Discover AI and expanded Defend AI to secure coding agents, productivity agents, and custom-built agent platforms. Agents are operating across enterprise systems with broad access, growing autonomy, and zero security oversight. That's why Straiker built Discover AI and Defend AI: to give security teams visibility into what agents are running and protection against what they might do.
Coding agents like Cursor, Claude Code, and GitHub Copilot are transforming how enterprise software gets built. With 85% of developers using AI coding tools, they have moved from experiment to essential infrastructure.
But this speed comes with serious risk. Coding agents ship more agents with minimal human oversight, creating vectors for endpoint takeover, data exfiltration, remote code execution, and tool manipulation through malicious MCP servers and agent skills. The risk extends to custom agents on AWS Bedrock AgentCore, Azure Foundry and Microsoft Copilot Studios, where tool calls chain across production systems and exposure compounds at every step. Enterprise productivity agents like Microsoft Copilot, ChatGPT Enterprise and Salesforce Agentforce compound this further, touching email, documents, CRM solutions and internal tools, often without security teams knowing which agents are active or what data they can reach. The agents aren't in danger; they are the danger.
Introducing Discover AI
Discover AI gives security teams complete visibility and governance over their AI agent landscape. Most organizations lack even a basic inventory of what agents exist, what they can access, or which MCP connections put them at risk. Discover AI changes that.
Discover AI provides:
- Agent discovery and inventory: Automatically identifies AI agents, MCPs and tools across coding tools, productivity platforms, and agent frameworks, giving security teams a centralized view of what agents exist, where they run and what systems they access.
- MCP vulnerability detection and enforcement: With over 12,000 MCP vulnerability databases, Discover AI detects vulnerabilities and risky configurations in MCP servers and connected tools, and can enforce controls to block malicious or compromised integrations.
- Security hygiene and misconfiguration detection: Flags agents operating in unsafe configurations such as unrestricted execution modes, excessive permissions or risky tool access.
- Agent observability and prompt classification: Classifies agent interactions by risk, giving security teams insight into what agents are doing and where threats are concentrated.
Defend AI: Runtime Security for Coding Agents and Agent Builder Platforms
Straiker's Defend AI is trained on millions of real-world agent traces from frontier AI labs and enterprise deployments, detecting agentic threats with sub-300ms latency and over 98% accuracy. With today's release, Defend AI expands with capabilities designed specifically for coding agents, productivity agents, and custom-built agent platforms:
- Runtime agent action tracing: Monitors agent actions and detects malicious instructions as they occur, identifying prompt injection, tool misuse, and unauthorized actions.
- Data exfiltration and destructive action prevention: Stops agents from leaking sensitive data or executing harmful commands before they impact enterprise systems.
- MCP and tool-chain risk detection: Identifies vulnerable or malicious MCP servers, tools, and agent integrations that can expose agents to supply chain and execution risks.
Defend AI can be deployed in monitoring mode via API integrations with Agent Builder platforms like Amazon Bedrock AgentCore, Azure Foundry and Microsoft Copilot Studio, allowing security teams to detect agent threats across their environment with minimal friction. For organizations requiring active enforcement, Defend AI also supports an inline gateway deployment that can block malicious agent actions in real time.
Together, Discover AI, Ascend AI and Defend AI make Straiker the agentic security control plane of choice, securing every category of AI agent across every stage of the security lifecycle, regardless of where agents run or how they were built.
Ken Buckler, CASP, research director, EMA, said: "Agentic AI is moving from experimentation to production at a pace that governance frameworks simply haven't caught up with. What's concerning is that our research shows nearly 80% of organizations have already deployed AI agents in production without formal policies in place to manage them. Most existing infrastructure was built for human users and traditional service accounts, not autonomous agents that can act, adapt and scale on their own. To manage this shift safely, organizations need to adopt a proactive approach that treats AI agents as first-class digital citizens with clear visibility, governance, and Zero Trust controls. We have witnessed attackers shifting from a 'break-in' to 'log-in' strategy, and the future of cyber threats is shifting towards politely asking an agentic AI for access."
David Levin, CISO, American Express Global Business Travel, said: "As an industry, we're rebuilding how we operate with AI agents at the center. Developers are already shipping with coding copilots, and that's just the beginning. As agents gain access to code, tools, and enterprise systems, the security stakes grow quickly. It's encouraging to see companies like Straiker focused on the protections needed to help enterprises adopt agents safely."
Scott Clinton, co-chair and co-founder, OWASP GenAI Security Project, said: "Agentic AI represents a major shift in how software operates, moving from AI user assistants toward multi-agent systems that can plan, act and interact autonomously across digital environments, introducing new rapidly evolving risks like behavior hijacking and privilege abuse. Addressing these challenges requires an open, community-driven approach to keep up. With contributions from organizations like Straiker, the OWASP Top 10 for Agentic Applications and related guidance were developed to give organizations clear, practical guidance on these emerging risks. As agentic systems rapidly move into production, open community resources help provide a holistic view of emerging threats and mitigations."
Meet Straiker at RSAC 2026
Expanded Defend AI and Discover AI are available now. Connect with Straiker at RSAC 2026, March 23–26 in San Francisco:
Early Stage Expo, Booth ESE-10 | Live demos
Agentic AI Leadership Breakfast, Tuesday morning | Register here
Private briefings and demos | straiker.ai/events/rsac
About Straiker
Straiker is the agentic security company. Straiker enables enterprises to discover their AI agent footprint, continuously test agent security through adversarial methods, and protect agents in production with runtime security. Founded by cybersecurity veterans, Straiker is trusted by leading enterprises to secure their AI transformation. Learn more at straiker.ai.
PR contact:
Shannon Van Every
Force4 Technology Communications
[email protected]
SOURCE Straiker
Share this article