Public-source research
Review what vendors say publicly, the exact quoted source text, the security requirement each statement may support, and what still needs verification.
Showing 361–380 of 1280 evidence records
Cisco claims runtime policies that detect and block unsafe agent actions, unauthorized tool usage, and privilege escalation.
Prevent harmful or unintended agent actions by enforcing policies on tool invocation, privilege levels, and action chains.
Cisco AI Defense materials reviewed did not provide a public claim for governing coding-agent commands, developer-workstation files or networks, integrated development environment (IDE) extensions, skills, hooks, secrets, or package actions.
Microsoft claims Agent 365 provides a centralized control plane to discover, manage, govern, and secure Microsoft and third-party AI agents.
Microsoft Agent 365 is a Microsoft 365 service that provides a centralized control plane for AI agents.
Microsoft provides open-source RAMPART and PyRIT tooling for repeatable adversarial and regression testing of agent and generative-AI systems.
RAMPART is an open-source testing framework that brings red teaming techniques directly into the development workflow.
Microsoft Defender for Cloud claims preview scanning of registered Azure Machine Learning models for embedded malware, unsafe operators, and exposed secrets before production.
AI model scanning provides proactive detection of unsafe or malicious artifacts and continuously monitors models for risk throughout the AI lifecycle.
Microsoft Foundry claims preview Model Context Protocol (MCP) governance by routing eligible tool traffic through an AI Gateway where Azure application programming interface (API) Management policies enforce authentication, rate limits, IP restrictions, and audit logging.
An AI gateway provides a single, governed entry point where you can enforce authentication, rate limits, IP restrictions, and audit logging
Microsoft Foundry claims automatic provisioning and lifecycle management of Entra agent identities, with Azure RBAC and token-based access for agent tool calls.
Microsoft Foundry automatically provisions and manages agent identities throughout the agent lifecycle.
Microsoft Defender for Endpoint claims runtime protection that can detect prompt injection and audit or block actions by supported local AI agents, including coding and CLI agents.
AI agent runtime protection helps you detect prompt injection at the device level and block or audit the agent's action
CrowdStrike claims unified AI visibility, governance, policy enforcement, and compliance logging across users, prompts, models, agents, Model Context Protocol (MCP) servers, endpoints, applications, gateways, and cloud environments.
AIDR maps relationships between users, prompts, models, agents, and MCP servers, and captures runtime logs for compliance.
CrowdStrike offers AI Red Team Services for organizations implementing generative-AI services and applications.
falcon-mcp and AI Red Team Services now available in new AI Agents and Tools category of AWS Marketplace.
CrowdStrike claims discovery of local models, Model Context Protocol (MCP) servers, integrated development environment (IDE) extensions, misconfigured AI tooling, vulnerabilities, and supply-chain risks on engineering endpoints before exploitation.
This helps security and engineering teams identify supply chain risks and misconfigured AI tooling before they become exploitable vulnerabilities.
CrowdStrike Falcon AIDR claims real-time protection for prompts, responses, agent actions, Model Context Protocol (MCP) servers, and AI/application programming interface (API) gateways with policy enforcement and automated response.
Falcon AIDR secures every prompt, response, and agent action in real time.
CrowdStrike claims access controls and policy enforcement across relationships among human and non-human identities, agents, models, Model Context Protocol (MCP) servers, and interactions.
AIDR provides unified visibility, real-time threat detection, data protection, access controls, and automated response.
CrowdStrike claims endpoint discovery and runtime protection for AI assets on engineering machines, including local models, Model Context Protocol (MCP) servers, integrated development environment (IDE) extensions, agent behavior, and sensitive-data flows.
AI assets that can be deployed on endpoints, especially by developers on engineering machines such as LLMs, MCP servers, and IDE extensions.
Wiz claims centralized AI inventory, posture, ownership context, compliance mapping, security baselines, prioritized risk, and remediation workflows across models, agents, services, data, infrastructure, and applications.
AI-SPM is designed to secure AI pipelines and accelerate AI adoption while maintaining protection against AI-related risks.
Wiz AI-SPM materials reviewed did not provide a public product claim for automated adversarial testing, model or agent red teaming, repeatable attack suites, or release-gate evaluation.
Wiz claims an AI bill of materials covering models, frameworks, dependencies, and libraries, plus model artifact scanning, repository and pipeline visibility, and attack-path analysis for model and training-data risk.
Analyze the components powering AI systems including models, frameworks, dependencies, and libraries.
Wiz claims runtime detection and response for prompt injection, rogue agents, malicious AI behavior, and Model Context Protocol (MCP)-connected AI systems using sensor and cloud telemetry.
Stop AI-native threats including prompt injection, rogue agents, and malicious AI behavior at runtime.
Wiz claims agent inventory and contextual mapping of agents to identities, workloads, data, access, capabilities, owners, and blast radius, with CIEM and secure-baseline checks.
Contextual Correlation: maps agents to the identities, workloads, and data they touch.
Wiz claims coding-agent and AI-integrated development environment (IDE) integrations that bring cloud and AI context into development, generate code fixes, and prevent cloud and AI risks from reaching production.
Meet your developers where they are and how they work with coding agent integrations that fix cloud and AI risks at the source.