AI runtime &amp; agent security

Automated vulnerability detection+5

the pioneering GenAI bug bounty platform designed to safeguard the future of artificial intelligence

A10 Networks

Network Detection & Response (NDR)

Flow-based anomaly detection+10

A10 Networks delivers secure, high-performance networking solutions that protect and scale critical applications across core, cloud, and edge environments

Agen.co

Runtime agent action governance+8

The same identity foundation, extended to the new workforce: AI agents. Identity-native governance for everything that acts on your systems.

Agentic Fabriq

Agent identity binding+10

The secure hub for agent identity, governance, and visibility. Control what your agents can access and do, for every user.

Air Security

Real-time prompt injection and jailbreak detection for LLM requests before they reach models used in agentic applications.

AIR secures every AI add-on before it becomes part of your enterprise. From discovery and continuous vetting to governance and runtime protection, AIR enables organizations to adopt AI safely. Air provides four solutions as a complete agentic security platform: - Air Control - Governance and control of agents across the enterprise, shadow ai discovery, posture management of agents configuration and connectivity - Air Defend - Runtime protection of agents behavior in real time. Visibility of every agent prompt and action, detection and response to dangerous and malicious behavior, runtime guardrails for agent behavior in realtime - Air Filter - Governing all agent add-ons which exist in the enterprise, continuously vetting of them, and detect and response to supply chain incidents. - Air Marketplace - Secured-by-default marketplace of pre-vetted add-ons, both external from open source and internally built, as a single source of truth for all enterprise usage

Akamai

Zero Trust Network Access as a serviceDevice posture based adaptive access+10

10 products

We make life better for billions of people, trillions of times a day

Akeyless

Secrets Management

Vaultless secrets management+11

Cloud-Native SaaS platform built to secure and manage every identity through a single pane of glass.

Alice

AI application risk testingRuntime guardrail enforcement+3

Keeping communicative tech safe, secure, and reliable.

AppOmni

SaaS Security Posture Management (SSPM)

Agentless SaaS configuration monitoring+11

AppOmni is positioned as a SaaS Security Posture Management (SSPM) platform, not a traditional multi-cloud CSPM vendor. The search results consistently clarify that AppOmni focuses on SaaS application security posture (Salesforce, Microsoft 365, ServiceNow, Google Workspace, Workday) rather than cloud infrastructure (AWS, Azure, GCP). While AppOmni uses CSPM terminology in legacy marketing, it does not compete in the infrastructure CSPM category. For cloud infrastructure posture management, AppOmni integrates with true CSPM solutions like Wiz. Organizations seeking multi-cloud IaC scanning, drift detection, and infrastructure misconfiguration remediation should evaluate dedicated CSPM vendors, not AppOmni.

AppSentinels Inc

Discover and inventory APIs+10

AppSentinels is a comprehensive API security platform that focuses on protecting the entire API lifecycle from development to runtime. It utilizes stateful API modeling and workflow analysis to identify sophisticated business logic attacks and data exfiltration that traditional WAFs often miss. The platform provides deep visibility into API discovery, vulnerability assessment, and real-time threat protection for REST, GraphQL, and AI-driven API endpoints.

Archestra.AI

Private MCP server registry+5

Archestra.AI is an open source security and governance layer positioned between large language models, AI agents, and enterprise data systems. The platform enforces a deterministic control layer so that access rules for agents stay fixed and predictable, independent of model judgment or prompt wording. It includes a private registry for Model Context Protocol (MCP) servers that governs which tools agents can reach, a retrieval augmented generation stack that runs inside the customer's own infrastructure, and a Kubernetes native orchestrator for multi-team deployments. The product is self-hostable and targets organizations connecting AI agents to systems such as HR platforms, email, and internal communication tools. Archestra.AI is a London based, early stage, seed funded company.

Astrix Security

Non-Human Identity Security (NHI)

Discover AI agents and NHIsMaintain a single AI inventory+9

Discover, secure, and deploy AI agents responsibly across the enterprise

Atsign

Encryption & Key Management

Edge-to-edge encryption+9

Atsign is a cryptographic identity and secure communications platform built around the atProtocol, with keys generated at the edge and encrypted data exchanged only between authorized Atsigns. In the Encryption & Key Management scope, its main value is non-custodial, peer-to-peer key handling: private keys stay on sender and receiver devices, and the infrastructure operator cannot decrypt customer data. It is best suited for IoT, distributed systems, and agentic AI teams that want end-to-end encrypted messaging without centralized key custody or exposed inbound ports. Atsign also offers adjacent secure remote access tooling, but the core security model is protocol-level encryption.

Aurva

Runtime monitoring for AI stack+9

Aurva provides a runtime security layer specifically designed for agentic AI architectures and LLM-driven workflows. It monitors granular data access patterns of AI agents, detects behavioral anomalies that signify prompt injection or agent hijacking, and enforces real-time compliance controls. The platform complements existing WAFs and API security tools by providing visibility into the internal logic and data orchestration of autonomous agents.

Autonomous Security

MCP traffic interceptionRuntime blocking of malicious skills+9

We build security and governance for AI agents running on desktops.

AvePoint

Data Security Posture Management (DSPM)

Sensitive data discovery and classification+11

AvePoint is the unifying Trust Layer for AI. AvePoint enables more than 28,000 organizations and 6,000 channel partners to protect, secure, and govern their entire AI estate across data, infrastructure, AI and agents for Microsoft, Google, Salesforce, and other leading cloud environments, so that enterprises can deploy AI with confidence and scale innovation without scaling risk.

Backslash Security

Application Security Posture Management (ASPM)

Internet-reachable code vulnerability detectionReachable package vulnerability analysis+10

The Security Platform for The New Agentic AI Fabric.

Barndoor AI

AI asset discovery and inventoryAI misconfiguration and exposure scanning+7

Barndoor is one platform for AI access, safety, and control.

Beyond Guard

AI agent discovery and inventory+8

BeyondGuard is the control plane for enterprise AI security. Inspecting every prompt, enforcing every policy, and auditing every decision your AI makes. AI security you can prove.

Beyond Identity

Multi-Factor Authentication (MFA)

Phishing-resistant passwordless MFA+11

Beyond Identity provides passwordless, phishing-resistant MFA built around device-bound cryptographic keys and device-native biometrics. Its MFA offering is aimed at organizations trying to replace passwords, push approvals, and OTPs with stronger authentication for workforce access. The platform uses an authenticator on endpoints and a cloud policy engine to verify both user identity and device trust at login. It is best suited for security teams that want MFA with continuous device posture checks and support for managed and unmanaged endpoints, while avoiding legacy second factors that can be phished or replayed.

Bifrost (by Maxim AI)

Unified multi-provider AI routing+11

At Maxim, we are building an enterprise-grade AI evaluation and observability platform to empower developers to ship their applications with quality, reliability, and speed.

Breeze Security

Attack Surface Management

External asset discoveryAttack path analysis+8

Breeze Security is a cyber asset attack surface management vendor focused on consolidating exposure data from existing security tools and mapping how misconfigurations chain into attack paths. In the ASM scope, it emphasizes continuous gap discovery, exposure correlation, and remediation prioritization across an organization’s security stack rather than standalone internet scanning. The product is best suited for security teams that already have multiple controls in place and need a single view of exposed weaknesses, asset gaps, and remediation order. Breeze also offers attack-path analysis and tailored playbooks as adjacent capabilities.

Britive

CIEM

Dynamic just-in-time cloud access+11

Britive provides dynamic Cloud Privileged Access Management (CPAM) and Cloud Infrastructure Entitlement Management (CIEM) with patented just-in-time (JIT) ephemeral access across AWS, multi-cloud, SaaS, hybrid, and on-prem environments. It enforces runtime identity access for human, agentic AI, and machine identities via a unified control plane, minting permissions only at execution and auto-destroying them post-task. Recognized by Gartner as a CIEM leader, Britive offers entitlement governance, anomaly detection, and SCIM-based synchronization with IdPs like Okta and Azure AD. Best for organizations needing granular, zero-standing-privilege controls in dynamic cloud ecosystems.

Burgus Security

Application Security (DAST/SAST)

Inspects LLM API traffic in production to enforce security checkpoints on prompts, responses, and agent-to-model exchanges for agentic applications.

I could not verify that a distinct vendor named Burgus Security has a documented Application Security product from the provided results. The only application-security-related result tied to the name is a directory-style entry that appears to refer to general application security services rather than a clearly described vendor platform. Based on the evidence available, Burgus Security cannot be reliably profiled as an AppSec product vendor, so the safest characterization is that its AppSec positioning is unconfirmed. If the intent was a different vendor name, the profile should be rebuilt from source documentation for that vendor.

CalypsoAI (F5)

Runtime prompt injection blocking+8

CalypsoAI, acquired by F5 in 2025 and rebranded as F5 AI Guardrails (formerly Inference Defend), delivers runtime security for AI models, agents, and applications. It combines agentic red-teaming with swarms generating 10,000+ new attack prompts monthly and real-time inference-layer protection against prompt injection, jailbreaks, data exfiltration, and denial-of-service. The platform includes the CalypsoAI Security Index and Agentic Warfare Resilience leaderboards for model benchmarking, plus policy enforcement for GDPR, HIPAA, and EU AI Act compliance. Best for enterprises deploying production AI copilots, RAG apps, and autonomous agents needing continuous vulnerability testing and defense integrated with F5 ADSP.

Cantina

Real-time prompt injection detection and blocking for AI agents running on Bedrock, OpenAI, and Azure OpenAI to prevent direct and indirect injection attacks

Cantina is the world's first truly agentic security platform: autonomous AI agents that don't just detect threats, but understand, respond, and adapt in real time.

Capsule Security

Enterprise AI agent runtime security+10

Capsule Security provides runtime security for AI agents. Using patented fine-tuned small language models, Capsule monitors autonomous AI agents in real time, detecting and stopping risky behavior before damage is done. With lightweight hook/API integration (no proxies, gateways, or SDKs), Capsule deploys in minutes and works across any agentic framework, including Coding Agents such as Cursor/Claude Code, and SaaS Agents such as Copilot Studio

Cequence Security

Complete API visibility and monitoring+10

Cequence Security sells API Security as part of its Unified API Protection platform, focusing on runtime discovery, inventory, compliance checks, and attack detection for internal, external, third-party, managed, unmanaged, shadow, and zombie APIs. It integrates with API gateways and reverse proxies to inspect live traffic and evaluate API usage and risk without adding client-side instrumentation. The product is best suited for enterprises that need continuous API attack-surface visibility and runtime protection across SaaS, on-premises, or hybrid environments. Cequence also offers adjacent bot management and WAAP capabilities, but its API Security scope centers on discovery, conformance, and blocking API abuse.

Certiv

Endpoint runtime agent discoveryPre-execution policy enforcement+10

We're building the control layer for the agentic workforce.

Cisco

CASBCisco SD-WAN integration+16

14 products

Cisco Umbrella is a cloud-delivered Security Service Edge (SSE) solution that enforces zero trust by continuously verifying identity, device posture, and context before granting access to applications. It converges multiple security functions, secure web gateway, firewall-as-a-service, cloud access security broker, and zero trust network access, into a unified cloud platform. Cisco Umbrella serves enterprises requiring distributed security across remote workers, branch offices, and on-premises infrastructure without complete network architecture overhauls.

Cline

Human-in-the-loop approval for actionsPlan and Act mode workflow control+9

Cline provides a secure, enterprise-grade execution environment for autonomous coding agents and LLM-driven development systems. It addresses the 'agentic security' gap by offering VPC and on-prem deployment options that ensure proprietary code and data never leave the corporate perimeter. The platform enables enterprises to govern agentic workflows with granular controls and IDE-agnostic integration, replacing unmanaged local AI browser extensions.

CloudSEK Research Pte. Ltd.

Threat Intelligence

Real-time threat intelligence monitoring+8

CloudSEK is a digital risk protection platform (DRPP) that utilizes AI to monitor the deep, dark, and open web for external threats. It provides automated detection of leaked credentials, brand impersonation, and exposed infrastructure to quantify digital risk. The platform complements internal SOC operations by providing an external-facing view of an organization's attack surface and supply chain vulnerabilities.

Clutch

Secrets Management

In-memory caching for secretsOpen-source Rust-based agent+5

Clutch is an identity security platform that discovers, maps, and governs non-human identities across cloud, SaaS, and AI environments. Using its Identity Lineage® graph, it connects AI agents, service accounts, API keys, OAuth applications, tokens, and secrets to the people, systems, and resources they interact with. This enables security teams to identify excessive privileges, detect identity-based risks, secure machine identities, and enforce governance across human and non-human access, helping organizations safely adopt AI at enterprise scale.

Coalfire

Security Operations

Managed security services+11

Coalfire's team of cyber legends who hunt, test, and neutralize vulnerabilities before they become headlines

Cranium

AI asset discovery and inventory+6

Born from KPMG Studio, Cranium gives enterprise teams one place to discover, observe, govern, secure and prove every AI and agentic system they build, buy or rely on, so innovation never outruns trust.

Crowsnest

Natural-language policy authoring+9

Crowsnest is the behavioral validation layer for enterprise security. Define your policies in natural language or as code; Crowsnest continuously validates applications, infrastructure, security tools, and the agents acting across them against the boundaries you've set - across pipeline, pre-production, and production, whether your SDLC is human-driven or agentic. Every finding ships as a plain-language explanation for executives and auditors alongside the technical evidence engineers need to remediate.

Cyera

Data Security Posture Management (DSPM)

Discover sensitive data across cloud and on-prem+10

Cyera is a pioneer in the data security space that empowers security leaders to discover their data attack surface, control the use of data, monitor, detect, and quickly remediate risk.

Darkhunt

compliance mappingguardrails+2

Darkhunt helps security teams discover, test, monitor and protect AI agents through continuous red teaming, observability and runtime protection, identifying risks including prompt injection, manipulation and unsafe agent behaviour.

Darktrace

Network Detection & Response (NDR)

Continuously monitors network traffic+11

Darktrace is a network detection and response vendor centered on self-learning behavioral analytics for north-south and east-west traffic. Its NDR product baselines normal activity for users, devices, and segments, then flags anomalous connections, command-and-control behavior, data transfer outliers, and other suspicious network patterns in on-premises, cloud, and hybrid environments. The platform is aimed at teams that need visibility beyond endpoint telemetry and want automated investigation and containment for network-level threats, including encrypted traffic and novel attack paths. Darktrace also sells adjacent security products, but this profile is limited to NDR capabilities.

DataKrypto

Continuous encryption for AI operations+9

DataKrypto provides deep-tech data protection specifically for LLM and AI workloads using Fully Homomorphic Encryption (FHE). The platform protects model weights from theft during inference and prevents data poisoning or prompt leakage by ensuring all operations occur in an encrypted state. It complements traditional encryption-at-rest by securing data-in-use within AI models and high-performance computing environments.

DeepKeep

Discovers AI modelsnotebooksdatasets+21

DeepKeep is a model agnostic AI security platform

Deliverance AI

Deliverance AI is positioned as a runtime security vendor for AI agents and large language model deployments. The company states its product provides guardrails that operate during execution to detect and block prompt injection attempts, unsafe or unauthorized actions, data leakage and policy violations before they affect production systems. It is aimed at organizations running LLM based applications or autonomous agents that need continuous runtime oversight rather than static pre deployment testing alone. No independent public confirmation of the company's website, product details, funding or customer base was found during this review, so this entry reflects only the positioning supplied for this catalogue assignment.

Drata

Compliance & GRC

Automated evidence collection for compliance frameworksContinuous control monitoring and audit readiness

Building the Trust Layer Between Great Companies

Dtex InTERCEPT

Insider Risk Management

Collects enterprise telemetry from users and devices

DTEX is the leader in risk-adaptive security, unifying human, data, and AI risk through a behavioral intelligence platform built for enterprise scale to detect threats early and prevent breaches.

Dynamo Ai

Automated AI risk evaluation and red-teaming+5

Dynamo AI specializes in the governance and security of Large Language Models (LLMs) and Generative AI agents. The platform provides automated red-teaming to identify prompt injections and jailbreaks, alongside real-time guardrails to prevent data leakage and ensure model alignment. It serves CISOs and AI leaders who need to maintain auditable compliance and safety across enterprise AI deployments.

Ember

Agentic SOC & Investigations

Autonomous triage and investigationCross-domain correlation for investigation+4

Security was built for the world before AI. We're building the one that comes after.

Endor Labs

Supply Chain Security

OSS dependency governance+11

Endor Labs is a software supply chain security platform focused on open source governance and CI/CD risk detection. The platform targets DevSecOps teams seeking to reduce alert fatigue while maintaining security posture. Endor Labs distinguishes itself through reachability analysis that eliminates up to 95% of false positives in vulnerability scanning, enabling developers to focus on exploitable risks. The vendor positions itself against traditional SCA tools by providing contextual intelligence on code usage patterns and safe upgrade paths.

Ericom Software

Browser & Web Isolation

Remote browser rendering in cloud containersEnd-to-end encrypted traffic inspection+10

Protect devices, apps, networks and data with isolation-based security solutions.

Eve Security

Block high-risk agent actions+7

Eve Security is an early-stage protection layer for AI agents, focusing on real-time policy enforcement and behavioral guarding. The platform monitors agent actions in production to prevent unintended behavior, prompt injection consequences, and policy violations. It acts as a runtime firewall for agentic AI, ensuring that autonomous systems operate within organizational safety boundaries.

F5 Advanced WAF

API endpoint discovery and inventoryGraphQL, REST, XML, and GWT protocol securityOWASP API Top 10 protection+9

We deliver and secure every app

Flint AI

Runtime prompt-injection detection+8

Flint AI is a local-first AgentOps CLI and platform focused on AI runtime and agent security, with controls aimed at pre-production testing and runtime policy enforcement for autonomous agents. It scans agent source code for risky tool access, missing guardrails, and misconfigurations, then runs adversarial evaluations against live agents to test prompt-injection and jailbreak resistance. Flint AI is best suited for teams building agents with frameworks such as LangChain, CrewAI, AutoGen, MCP, or OpenAI/Anthropic SDKs that want validation and enforcement before deployment.

Fortanix DSM (Data Security Manager)

Encryption & Key Management

Centralized enterprise key management+11

Fortanix is a data-first security company and a pioneer in Confidential Computing. We help enterprises discover, assess, and remediate data exposure risks across hybrid multicloud environments to maintain the privacy and compliance of their most sensitive and regulated data, wherever it may be.

General Analysis

AI asset discovery & inventoryAutomated AI red-teamingPrompt injection testing+3

General Analysis is a security platform for organizations that run AI agents and large language model systems in production. It discovers and inventories AI assets, including models, vector stores, MCP servers and agent workflows, by connecting to code repositories, LLM providers and cloud infrastructure, and scores each asset by risk. The platform runs automated red team simulations covering prompt injection, tool misuse, sensitive data retrieval and multi step exploit chains to find exploitable gaps before an incident occurs. It also deploys runtime guardrails that block threats and monitor for policy violations and performance drift. Buyers include enterprise security teams securing employee copilots, customer support agents and AI workflows in healthcare, legal and financial services.

Geordie

Real-time contextual agent governanceBehavioral baselining and anomaly detectionDeterministic real-time agent remediation+8

Lighting the way to safe and scalable enterprise agent adoption

HOL

local-firstpolicy enforcement+2

HOL Guard is a local-first runtime control / open-source antivirus for AI agents. Apache-2.0. It enforces policy before a risky agent action runs on the machine. Product site: https://hol.org/guard

JetStream Security

AI visibility and discovery+11

JetStream Security is a security-first AI governance platform that fits the AI Runtime & Agent Security category by governing live agent and LLM activity with identity-bound controls, runtime policy enforcement, and MCP oversight. Its public materials emphasize continuous discovery, blueprint-based control, and runtime governance for agents, tools, workflows, and model traffic rather than static AI inventory or posture scanning. It appears best suited for enterprises running production agentic systems that need to verify third-party MCP servers, restrict tool permissions, and trace every invocation to an accountable owner. Adjacent capabilities like AI spend tracking are present but secondary in this profile.

Keycard Labs, Inc.

Short-lived task-scoped agent credentials+11

Keycard Labs provides an authentication and authorization framework specifically designed for AI agents and LLM-driven applications. It replaces manual, hard-coded authorization logic with a centralized control plane for managing agent identities, permissions, and session visibility. The platform captures a full audit trail of agentic actions, allowing CISOs to enforce granular access controls and ensure that autonomous agents operate within defined security boundaries.

Lakera (Check Point)

Block prompt injection in real time+8

Check Point Software Technologies is a global leader in cyber security solutions, dedicated to protecting corporate enterprises and governments worldwide.

Manifold Security

Agentless endpoint runtime visibilityBehavioral monitoring of agent actions+3

AI Detection and Response, purpose-built for the agentic enterprise.

MATVIS

AI inventory managementFull-stack attack path analysis+4

Enterprise AI is moving fast. The security infrastructure isn't keeping up.

Menlo Security

Browser & Web Isolation

10 products

As a pioneer of browser security, Menlo Security delivers a solution with a comprehensive approach to enterprise browser security, protecting users where they work and securing applications from internet-borne attacks.

Cloud-based remote browser isolation+11

Metomic

Data Loss Prevention (DLP)

Sensitive data discovery across SaaS apps+9

Metomic is AI data security for enterprise SaaS, the control plane for sensitive data across every SaaS tool and every AI surface. We find it, fix it, and prove it. Content-level scanning inside Slack, Jira, Confluence, Drive, SharePoint, Salesforce, and more. Automated remediation that revokes sharing, redacts PII, notifies owners, and coaches users at the point of risk. Metomic Cleanser delivers a redacted copy of sensitive content that's safe for an LLM to consume.

Mimic

Cyber Resilience & Recovery

Backup server continuity+8

Mimic is a security vendor whose backup and disaster recovery offering centers on automated recovery of business-critical applications and data into a clean environment within 24 hours. Its published recovery page emphasizes restoring applications and configurations within hours and avoiding ransom payment, which places it closer to cyber recovery than basic file backup. It appears best suited for organizations that need fast restoration of critical workloads after ransomware or destructive incidents. Publicly available material is limited, so the exact depth of platform coverage and deployment model is not fully documented in the provided sources.

Mindgard

Real-time AI threat detection and response+10

Mindgard gives organizations the visibility and protection to discover, assess, and defend their AI systems and agents.

Mint Security

MCP gateway governance+11

Mint Security is an AI agent security platform focused on runtime governance for agentic systems and MCP-connected tools. It sits in the AI Runtime & Agent Security category by enforcing policies at execution time, tracing tool calls, and blocking unsafe agent actions rather than doing pre-deployment inventory or posture scanning. The vendor positions the product for security and platform teams that need centralized control over AI agents, MCP servers, and agent workflows in production. Mint also offers adjacent MCP governance capabilities, but the runtime control plane is the core security use case.

Mitiga

Cloud Detection & Response (CDR)

Panoramic AwarenessAttack Decoding+2

Mitiga is a cloud detection and response vendor focused on detecting, investigating, and containing active threats across cloud, SaaS, identity, and AI environments. Its CDR platform emphasizes zero-impact breach prevention, with cloud-native telemetry analysis, attack reconstruction, and guided response for SecOps teams that need visibility beyond CNAPP and posture tools. Mitiga positions itself for enterprises with distributed, multi-cloud and SaaS-heavy footprints that need forensic context and rapid triage without requiring deep cloud expertise.

NeuralTrust

Runtime agent protection+5

NeuralTrust is an AI security platform that discovers and protects autonomous AI agents, models, and tools across an enterprise. Its runtime layer inspects every request an agent makes to models, tools, MCP servers, and data before execution, blocking prompt injection, data leakage, and unauthorized API calls in real time. The platform combines an agent gateway that enforces unified policies across homegrown apps, SaaS tools, and platforms like ChatGPT, Gemini, and Copilot, an agent discovery and posture module that inventories deployed agents and workflows, and adversarial red teaming to surface model vulnerabilities before deployment. It is built for enterprise security and compliance teams and deploys as SaaS, on premises, or hybrid, keeping the data plane local for data sovereignty.

NewCore

Identity & Access Management (IAM)

Identity discovery and unification+11

NewCore is a workforce identity platform positioned for IAM use cases across humans and non-human identities. Its published materials emphasize identity discovery, SSO, MFA, lifecycle management, directory services, and inline policy enforcement, with support for deploying alongside an existing IdP rather than replacing it. It appears best suited to enterprise security and identity teams that need to unify access control for employees and AI agents, especially where browser-based authentication, agent governance, and token-based access control are in scope.

Nokod Security

Agent discovery and ownership mapping+4

Nokod Security is a security vendor focused on runtime protection for AI agents and citizen-built automations, especially in no-code and low-code environments. Its Adaptive Agent Security product provides live visibility, behavioral baselining, and policy enforcement across the Agent Development Lifecycle, with emphasis on stopping risky tool use and unauthorized data access as agents run. It is best suited for enterprises using Microsoft Copilot Studio, Power Platform, ServiceNow, UiPath, and similar agentic platforms that need inline controls rather than only post-deployment review.

Nord Security

Cloud-delivered secure access architecture+10

Nord Security offers a suite of business tools including NordLayer for network access and NordPass for credential management, focused on the SMB and mid-market segments. It provides a secure service edge (SSE) approach to remote access, replacing legacy VPNs with a Zero Trust Network Access (ZTNA) model. The platform integrates identity-centric access control with password security and threat exposure monitoring.

Novee

Autonomous AI red teaming+8

Novee is primarily an AI penetration-testing vendor, not a pure AI runtime control plane. For AI Runtime & Agent Security, it is best understood as an attack-simulation and validation product that probes LLM apps, copilots, and agents for prompt injection, jailbreaks, tool abuse, and workflow manipulation, then validates exploit paths and remediation. It is most suitable for teams that want adversarial testing of agent behavior across OpenAI, Anthropic, and open-source stacks before deployment or during continuous security validation.

Oligo Security

Cloud Workload Protection (CWPP)

Runtime workload protection+8

Oligo Security is primarily a runtime security vendor, not a native CSPM specialist. In cloud security evaluations, it is best understood as a platform for detecting and blocking active exploitation in cloud workloads, with emphasis on runtime context rather than posture scanning or misconfiguration management. Its cloud-security materials focus on protecting modern applications, cloud workloads, and AI systems at execution time, which makes it a fit for teams that want runtime threat detection and exploit prevention alongside other cloud security controls.

Omada

Identity Governance & Administration (IGA)

Identity lifecycle management+11

Omada Identity provides full-featured Identity Governance and Administration (IGA) solutions, including on-premises and SaaS (Omada Identity Cloud) deployments for managing identity lifecycles, access requests, certifications, and compliance reporting. It supports hybrid/multi-cloud environments, ITSM integrations, and granular visibility into user access patterns. Key capabilities include risk-based access control, account reconciliation, and Segregation of Duties (SoD) policy enforcement. Best suited for enterprises requiring policy-driven governance to handle complex workflows, mitigate risks, and ensure regulatory compliance without multiple disparate tools.

Openlayer

Runtime prompt injection detection and blocking+9

Openlayer is the AI governance platform that helps enterprises discover, test, monitor, govern, and optimize AI systems across their entire lifecycle, from prototype to production.

Operant Networks

Comprehensive AI Security Index for model resilience

Operant is built by a world-class team with decades of experience in networking, machine communications, and cybersecurity. Our founders have a track record of commercializing next-generation technologies, often creating entirely new product categories. Today, we are applying that expertise to the next frontier: trust for machine and AI communications in dynamic, distributed environments. Recognized by the U.S. Department of Energy as building "game changing" technology, Operant is pioneering Zero Trust for machines and AI, proven in critical infrastructure, and built to extend as AI reaches further into the physical world.

P0 Security

CIEM

Identity inventory and entitlement discovery+7

P0 Security provides a cloud identity security platform that focuses on eliminating standing privileges through JIT (Just-in-Time) access and least-privilege enforcement. The solution automates the governance of human, workload, and AI agent identities across multi-cloud environments, ensuring that high-risk access is ephemeral and strictly vetted. It replaces traditional static PAM for cloud environments and complements CSPM by securing the identity layer.

Palo Alto Networks

Firewall / NGFW

21 products

Palo Alto Networks is a major vendor in the Firewall / NGFW market, best known for PAN-OS-based next-generation firewalls and Cloud NGFW. Its firewalls combine application-aware policy, user-based controls, and threat prevention to inspect traffic, including encrypted sessions, and block known and unknown threats. It fits enterprises that need granular segmentation, internet edge protection, and consistent policy across physical and cloud deployments. Adjacent offerings exist, but the core firewall line remains centered on network traffic control, inspection, and prevention.

Active-passive firewall high availability+22

Pillar Security

Application Security (DAST/SAST)

Pillar Security provides an AI security platform designed to discover, govern, and secure AI agents across an organization. It offers capabilities to map AI landscapes, assess risks, red team AI systems, enforce data policies, and apply adaptive runtime guardrails for AI applications, models, and agents. The platform aims to ensure compliance and provide real-time protection against AI-specific threats.

AI Discovery & Posture+3

PrimeSec Inc.

Static application security testingDynamic application security testing+2

Agentic development has made shipping secure products harder than ever. We built Prime to be the trusted advisor that knows your architecture.

Promptfoo

Automated red-teaming for AI applicationsDynamic adaptive scan generationApplication-focused vulnerability testing+6

Promptfoo helps developers and enterprises build secure, reliable AI applications.

Prompt Security (SentinelOne)

Block prompt injection in AI interactions+11

Prompt Security delivers runtime GenAI security across employee AI tools, custom LLM applications and AI agents. Acquired by SentinelOne in 2025, the platform is being woven into SentinelOne's Singularity stack to give SOC teams unified visibility and control over shadow AI usage, prompt injection attempts and sensitive data exposure across the enterprise. Best suited for security teams that need centralized governance of GenAI usage with inline guardrails, DLP and audit trails across browsers, IDEs and homegrown AI applications.

QuilrAI

Detects prompt injection and jailbreak attempts+10

QuilrAI offers an autonomous decision engine designed to secure interactions between human users and AI agents. By analyzing content, context, and intent in real-time, the platform identifies and blocks malicious actions within agentic workflows. It provides point-of-decision education to users, effectively shielding enterprises from breaches originating in LLM-driven or autonomous agent environments.

Raven

Application Security Posture Management (ASPM)

Runtime exploit prevention+5

Raven.io provides a runtime application security platform that runs inside cloud-deployed applications to detect and block malicious code execution as it happens, independent of whether a CVE exists. The platform includes application detection and response for forensic visibility down to libraries, functions and call paths, a runtime software composition analysis capability that uses reachability data to de-prioritize vulnerabilities, and a module for discovering and controlling AI agents operating inside application environments. It is aimed at enterprises running internet-facing applications on Kubernetes, containers and compute instances across AWS, GCP and Azure, supporting languages including Java, Python, JavaScript, Go, Ruby, PHP, C++ and .NET. Customers include organizations in regulated industries such as insurance and financial services.

Realm Labs

Continuously monitor AI behavior at runtime+5

Realm Labs provides an AI Trust and Security platform designed to monitor and secure LLM applications in production environments. It focuses on identifying and mitigating 'surprises' such as model hallucinations, prompt injections, and data leakage by providing a transparency layer for enterprise AI workflows. The solution complements existing LLM development frameworks by adding a dedicated security and observability shim to ensure model outputs align with corporate policy.

Reco

Runtime prompt and tool-call inspection+9

Reco is the leader in Agentic Ecosystem Security, securing everything around the agent, not just the prompt.

Relyance AI

Data Security Posture Management (DSPM)

Continuous sensitive data discovery+10

To secure the future of data-driven enterprises by making every data journey visible, controlled, and compliant, from the first line of code to the final AI output.

Salt Security

Enrich API intelligenceDetect API attackers earlyProactive API posture improvement+3

The world leader in Agentic Security

Sector8 AI

Identity Governance & Administration (IGA)

Sector8 places an independent decision gate between agents and real-world actions.

SecurEnds

Automated identity lifecycle management+11

Govern every identity across your enterprise. Automate user access reviews, enforce least privilege, and simplify compliance.

Securiti

Privacy & Consent Management

Website cookie scanning and consent banner integration

At Securiti, our mission is to enable organizations to safely harness the incredible power of Data & AI.

Semgrep

Static Application Security Testing (SAST)

Static application security testing+11

Semgrep is a developer-focused SAST platform that combines static analysis with multimodal AI reasoning to detect vulnerabilities in source code. The platform unifies SAST, SCA, and secrets scanning, emphasizing reduction of false positives through code context and prior decision patterns. Semgrep integrates with CI/CD pipelines for continuous scanning and correlates findings with dynamic testing via StackHawk's DAST to validate exploitability. Best suited for development teams prioritizing early vulnerability detection with minimal alert noise.

Silverfort

Identity Threat Detection & Response (ITDR)

10 products

Silverfort secures every dimension of identity. We break down the silos of identity infrastructure and point solutions to eliminate security gaps and blind spots once and for all. The result? Identity security without limits, that doesn't slow down the business.

Monitor authentication across hybrid environments

Skyrelis

Agent Identity ModelLeast-Privilege Scoping+6

Skyrelis is a specialized security platform providing a runtime policy layer for autonomous AI agents, ensuring safe production deployment. It enables enterprises to enforce security and compliance policies on agentic AI workflows in real-time without requiring modifications to the underlying agent code. The platform manages risks associated with autonomous decision-making across various regulatory frameworks and geographic boundaries.

Sonar

Static Application Security Testing (SAST)

Source code vulnerability scanningSupport for 40 plus languages+7

Sonar helps developers deliver high quality and secure software by analyzing code they write, AI-generated code, and code leveraged from third parties (like open source libraries). Sonar's integrated approach to improving code quality and code security catches these issues before they make it into production, helping developers reduce technical debt and code complexity over time.

SOVEREIGN AI by ARMOR

Governed AI request gateway+7

Sovereign AI is built by Armor. For over a decade we have secured the world's most important data and driven painless compliance outcomes for thousands of customers across forty countries, against the highest bars in the business: HIPAA, PCI DSS, GDPR, HITRUST.

SplxAI

Real-time prompt injection and jailbreak detectionInput and output guardrail enforcement+10

SPLX helps global enterprises secure AI systems end-to-end, with scalable red teaming, real-time threat protection, and automated AI compliance and governance.

StealthMail

Email Security

Replaces standard email content transmission with authenticated 'Stealth-Links' that contain no user information, bypassing SMTP vulnerabilities and preventing intermediaries from accessing message bodies or attachments

Email needs to earn the right to be trusted by millions of people and organizations with their most sensitive information. We see a future where everyone is confident that their private correspondence is genuinely secure.

Straiker

AI agent inventory discovery+6

Straiker is an agentic AI security vendor whose AI-SPM offering centers on discovery and posture management for AI agents rather than general cloud posture. Its Discover AI product inventories AI agents, MCP servers, tools, and integrations, then surfaces misconfigurations, risky permissions, and unsafe connections across builder platforms and coding agents. Straiker is best suited for security teams that need visibility into agent sprawl and exposure in environments using Bedrock AgentCore, Azure Foundry, Copilot Studio, Cursor, Claude Code, or GitHub Copilot. The company also sells adjacent runtime protection and red-team modules, but those are outside AI-SPM scope.

Synopsys

Application Security (DAST/SAST)

11 products

Synopsys is the leader in engineering solutions from silicon to systems, enabling customers to rapidly innovate AI-powered products.

Static application security testing for source code

Tego AI

AI agent discovery & visibility Real-time agent monitoring Granular runtime access control Purpose- & intent-aware security MCP gateway security & enforcement Local and cloud agent coverage UsersData & Tools risk analysis Dynamic policy enforcement Sensitive data exposure detection Agent posture & risk management Inline security architecture Centralized AI agent governance

Security built for the agentic era, by the people who built security for the cloud era.

Tenet Security

Agent-side action simulation+5

Tenet Security provides runtime protection for autonomous AI agents inside enterprises. Its platform, built around a technique it calls Agent-side Simulation, models an agent's likely next actions before they execute against live systems and can block actions judged risky before they occur, rather than alerting after the fact. It addresses threats specific to agentic AI, including unauthorized access, data exfiltration, agent manipulation and a technique the company calls Agentjacking, where malicious content hidden in emails, documents, logs or databases covertly alters agent behavior. The company was founded by former Cisco AI Defense researchers Barak Sternberg and Nevo Poran and emerged from stealth in 2026 with a six million dollar seed round.

Teramind

Insider Risk Management

Behavioral analytics for data loss detectionContent inspection and data classificationEndpoint and network data monitoring+9

User activity. Insider risk. One platform.

ThreatDown

Endpoint Detection & Response (EDR)

Advanced behavioral threat detection+5

ThreatDown is redefining cybersecurity for businesses of all sizes. We strip away the bloat, the cost, and the confusion, replacing it with powerful, intuitive security that protects thousands of organizations worldwide from the most advanced threats.

Tigera Calico

Container Security / CNAPP

Container image assurance scanningRuntime threat defense for workloads+10

Secure your Kubernetes workloads. Govern your AI agents.

Transmit Security

Multi-Factor Authentication (MFA)

Passwordless MFA with passkeys+11

Transmit Security is an enterprise identity vendor whose MFA capabilities are delivered through its Mosaic platform, which combines passwordless authentication, biometrics, and step-up controls for customer-facing logins. In the MFA category, it is best known for FIDO-based authentication and app-less biometric methods that reduce password dependence while supporting high-assurance access. Its core buyers are large enterprises in regulated sectors such as banking, insurance, and retail that need strong customer authentication across web and mobile channels. The company also sells adjacent CIAM and fraud-prevention capabilities, but those are outside this profile’s scope.

Trent AI

Continuous code and infra scanning+5

Trent AI builds an AI-native security platform for organizations building or operating autonomous AI agents. The product uses a coordinated set of scanning, judgment, mitigation and evaluation agents to continuously assess code, infrastructure and AI agent runtime behavior, detecting risks such as prompt injection, tool misuse, unintended actions, data exfiltration and privilege escalation. It prioritizes findings by real-world exploitability, generates remediation fixes or prompts for developer tools, verifies that fixes are deployed, and maps controls to compliance frameworks such as SOC 2 and NIST. It targets AI-native startups without dedicated security staff and enterprise security teams seeking automation, and can be deployed in public cloud, a customer VPC, or on-premises using frontier, open-source or private AI models.

TrojAI

AI asset discovery and inventory+4

TrojAI is a Canadian AI security vendor focused on securing AI models, applications, and agents across build-time and runtime, with a separate emphasis on AI model testing and policy enforcement. Within AI Security Posture Management, its Detect capability is the relevant fit: it red teams models before deployment to surface behavioral risks and remediation guidance. The vendor is a stronger fit for enterprises that need pre-deployment AI risk assessment and controls around model behavior rather than a pure inventory-only posture tool. Adjacent runtime defense products exist, but they are outside this AI-SPM scope.

Turbot

Cloud Security / CSPM

Enforce cloud guardrails in real time+7

Turbot offers an automated cloud governance and remediation platform that enforces real-time guardrails across AWS, Azure, and GCP. It differentiates itself from traditional CSPMs by moving beyond simple alerting to active prevention and auto-remediation of misconfigurations. The platform manages cloud infrastructure drift and ensures compliance with frameworks like CIS, NIST, and HIPAA through programmable policy-as-code.

Tyk

Authentication and authorization enforcement+11

Tyk is an open-source API gateway and API management platform that, in the API Security scope, provides request authentication, authorization, traffic controls, and policy enforcement for REST, GraphQL, gRPC, TCP, and SOAP APIs. It is best suited to teams that want to secure and govern APIs at the gateway layer while using a self-managed or cloud deployment model. Tyk also includes adjacent API management components such as analytics and a developer portal, but its security value is centered on controlling access, transport security, and request filtering at the edge.

Unbound

Discover AI coding agents and MCP servers+7

Unbound provides a security and governance layer for AI coding assistants and autonomous agents. It specializes in detecting the use of AI tools, MCP (Model Context Protocol) servers, and agentic workflows to prevent insecure code generation and data leakage. The platform allows organizations to apply fine-grained controls over what AI agents can access and the code they are allowed to submit to repositories.

Upstream Security

Runtime AI and API security+11

Upstream Security is primarily an automotive and physical-AI security vendor, but its AI runtime offering extends into agent and API enforcement through its Runtime AI and API Security platform. In this scope, it monitors traffic across AI and API ecosystems, discovers agents and endpoints, and applies stateful inspection and custom detections for OWASP MCP and LLM risks, prompt-injection-style abuse, and business-logic misuse. It is best suited for organizations that need runtime controls around agentic workflows and API-backed AI services, especially in connected-vehicle and industrial environments.

Verno Labs

Adversarial testing for AI agents+4

We're building the offensive AI that the next decade of enterprise security will depend on, and putting it in the hands of the companies who can't afford to wait.

Vijil

Runtime prompt injection and jailbreak defense+4

Vijil is the trust infrastructure that enterprises need to use AI agents with reliability, security, and safety.

Vorlon

Maps live data flows across agents+11

The Agentic Ecosystem Security Platform. Powered by DataMatrix™, Vorlon's patented intelligent simulation engine, the platform monitors every agent action, detects threats across the full integration layer, and enforces data security in real time across every system AI agents touch.

Votal AI

Continuous AI model discovery+6

Secure every AI interaction, from prompt to tool call, guardrails, identity, data policies, and tool authorization in real time.

Vyper Security

runtime prompt injection blocking+11

Vyper Security is a vendor in the AI Runtime & Agent Security category that focuses on protecting production LLM and agent workflows during execution, where prompts, tool calls, and model outputs can be manipulated. Based on publicly available material, it appears to position around runtime enforcement rather than AI inventory or model posture scanning, which keeps it aligned with agent execution control. It is best suited for security teams that need to gate agent actions, inspect model interactions, and reduce prompt-injection and unsafe tool-use risk in live AI applications.

Wider Security

Security Operations

Security engineeringCloud security+8

Wider Security is a veteran-owned cybersecurity services firm founded in 2019 that provides security operations-related support through engineering, integration, cloud security, and risk management work. In the Security Operations category, it appears to be a services-led provider rather than a software platform, with emphasis on securing government and defense environments and supporting DoD IT networks. It is best suited for organizations that need hands-on security operations support, systems integration, and technically skilled staffing rather than a standalone SOC product.

Willow

Agent identity and access controlRuntime guardrail enforcement+6

Willow builds the identity and access layer for AI agents at work.

Wing Security

SaaS Security Posture Management (SSPM)

SaaS application security posture assessmentAttack surface minimization for critical applications+7

Agent-based security for agent-based risk. By With Wings.

WitnessAI

Blocks prompt injections at runtime+11

WitnessAI is an enterprise-grade AI security platform that focuses on runtime governance and network-level visibility for Large Language Models (LLMs) and shadow AI. It provides an 'AI Firewall' layer that monitors interactions between users and AI tools to prevent data leakage and enforce intent-based usage policies. The solution helps CISOs enable safe AI adoption by intercepting prompts and responses for compliance and security violations in real-time.

WSO2 API Manager

OAuth2 API access control+11

WSO2 API Manager is an open-source API management platform that, in the API Security category, centers on gateway-enforced authentication, authorization, throttling, threat protection, and traffic mediation for HTTP APIs and, in newer releases, GraphQL and asynchronous APIs. It is aimed at enterprises that need policy-driven control over exposed APIs across cloud, hybrid, and on-prem deployments. WSO2 also positions it as part of a broader API management stack, but its security value here is the gateway and policy enforcement layer rather than endpoint scanning or runtime app protection.

Xage Security

OT & ICS Security

Layered multi-factor authentication+10

Xage Security is a zero-trust access vendor whose IAM offering centers on identity-based access for mixed IT, OT, and edge environments. In the IAM scope, it provides multi-factor authentication, federation, single sign-on, and policy-based access orchestration across multiple identity providers and local directories. Xage is best suited for industrial, critical infrastructure, and distributed enterprise buyers that need access control across legacy systems, remote sites, and intermittently connected environments. The company also sells adjacent zero-trust and privileged access capabilities, but its IAM value is rooted in layered identity enforcement.

Xeris AI

Super AI Agent runtime governance+8

Xeris AI provides specialized security for Agentic AI and workflows utilizing the Model Context Protocol (MCP). The platform uses autonomous 'Super Agents' to monitor, govern, and intercept AI behaviors at the agentic layer, preventing unauthorized actions before they execute. It is designed to tackle the unique challenges of autonomous AI agents that can perform tasks and access data independently within an enterprise environment.

Xiid Corporation

Outbound-only AI runtime isolation+10

Xiid Corporation is a security vendor whose AI Runtime & Agent Security story centers on its Terniion architecture and SealedTunnel/SealedChannel controls for constraining AI and agent traffic at runtime. In this category, it appears strongest on network-mediated enforcement, MCP tool-call logging, and tamper-evident controls rather than model inventory or posture scanning. Its material is oriented toward organizations that need to keep agentic workflows behind zero-trust boundaries and inspect or restrict agent operations while preserving cryptographic auditability. It is best suited for buyers looking to secure AI execution paths, tool usage, and privileged access around deployed agents.

Zenity

AI agent discovery and inventory across clouds+5

Zenity provides unified security and governance for AI agents and Low-Code/No-Code applications across the enterprise ecosystem. It enables organizations to discover shadow AI, assess risk posture, and enforce security policies on GenAI agents such as Copilots. The platform offers full-lifecycle protection from development discovery to inline runtime response, addressing the unique risks of AI-driven automation.

Zentera Systems, Inc.

Zentera Systems offers a Zero Trust Network Access (ZTNA) and microsegmentation platform centered around its Virtual Chamber technology. It provides a secure overlay for IT, OT, and AI infrastructure that requires no changes to underlying network architecture. It is often used to replace traditional VPNs and complex firewall-based segmentation with a more agile identity-centric access model.

Zero Touch Deployment+9

ZeroPort