Public-source research
Vendor evidence
Review what vendors say publicly, the exact quoted source text, the security requirement each statement may support, and what still needs verification.
Research library coverageCounts describe the research workflow, not vendor quality or product effectiveness.View details
Evidence records1280
Source checked863
Needs verification0
Source captured0
No supporting claim found411
Excluded from evidence6
Palo Alto Networks says Prisma AIRS with Portkey is intended to monitor, govern, and protect AI applications, models, and autonomous agents.
With the integration of Portkey’s AI Gateway into Prisma AIRS, we will deliver a centralized control plane to monitor, govern, and protect every AI application, model, and autonomous agent across your organization
Related framework references (5)
Cisco AI Defense mentions red teaming AI models and agents, but reviewed materials do not show broad autonomous-agent telemetry or graph visibility.
Try AI red teaming for your models and agents today
Related framework references (5)
Microsoft says supported AI agents inherit the same Purview security and compliance capabilities as their parent AI app.
Where these AI apps support agents, they inherit the same security and compliance capabilities as their parent AI app.
Related framework references (5)
Noma claims discovery provides visibility and context across models, agents, Model Context Protocol (MCP) servers, data sources, and their dependency chains.
Discovery provides visibility and deep context for your entire AI landscape: every model, every agent, every MCP server, every data source, and crucially, how they all connect.
Related framework references (5)
JetStream claims Design Control maps agents, models, tools, datasets, and identities and how they interact, then compares intended design to runtime reality as systems evolve.
Design Control turns raw discovery into an approved design by mapping how your AI actually works—agents, models, tools, datasets, and identities—and how they interact. It becomes a living operational contract with versioning and change control, so teams can document intent and compare it to reality as systems evolve.
Related framework references (5)
Netskope claims AI Command Center provides visibility across generative AI apps and autonomous agents.
Gain comprehensive visibility across your entire AI environment, from genAI apps to autonomous agents
Related framework references (5)
Singulr Agent Pulse claims to discover every AI agent and build context graphs of tool connections, data access, Model Context Protocol (MCP) servers, and permissions.
Discover every AI agent in your environment on any platform. We create the context graph of tool connections, data access, MCP servers, and permissions to show how agents interact with your enterprise systems.
Related framework references (5)
SentinelOne Prompt Security claims searchable audit logs for every agent action, decision, and enterprise system interaction.
Get a searchable audit log of every agent action, decision, and enterprise system interaction
Related framework references (5)
AIM Security (Cato) claims it monitors every interaction between agents, models, and Model Context Protocol (MCP) servers to keep agents secure and compliant.
monitor every interaction between agents, models, and MCP servers, to ensure agents operate securely, remain complaint, and align with business needs.
Related framework references (5)
Oasis Security claims its agentic access platform captures every session and shows each ephemeral identity, granted access, and real-time activity.
Capture every session: intent, policy, identity, activity, and expiration, for total visibility and compliance.
Related framework references (5)
Grip Security claims visibility into AI tools, agents, software as a service (SaaS) apps, identities, and exposure points.
what those tools and agents are doing, where exposure and risk exists, and how to take action.
Related framework references (5)
WitnessAI claims it discovers running agents and the external Model Context Protocol (MCP) servers and tools they connect to, and governs agent actions with runtime security.
Discover which agents are running and what external MCP servers and tools they connect to
Related framework references (5)
Lasso Security claims it maps models, system prompts, tools, and guardrails and provides full context on attacks including which agent was targeted.
Maps models, system prompts, tools and guardrails
Related framework references (5)
Harmonic Security claims it governs AI interactions across employees and agents, including browser, desktop, agent, and Model Context Protocol (MCP) surfaces.
Every interaction visible. Every interaction governable.
Related framework references (5)
Reco claims it maps every agent in the environment to its owner, permissions, and risk with lineage and context.
Know exactly what every agent in your environment can access, who owns it, and where the risk is. Before the next class of AI finds out first.
Related framework references (5)
Lakera claims AI-agent landscape discovery, risk assessment, and real-time protection enforcement.
Discover your agent landscape, assess risk, and enforce protection in real time.
Related framework references (5)
Zenity claims it monitors step-level agent execution and enforces inline controls to stop unsafe actions.
Monitor step-level agent execution, correlate behavior with context, and enforce inline controls to stop unsafe actions before they impact the business
Related framework references (5)
Nightfall claims unknown Model Context Protocol (MCP) servers can be discovered across endpoints and surfaced in a dashboard with policy defined to block Model Context Protocol (MCP) servers.
Unknown MCP servers discovered running locally across 12 endpoints > full inventory surfaced in the dashboard. Policy defined to block MCP servers.
Related framework references (5)
LayerX claims monitoring and protection for agentic AI browsers, embedded browser agents, AI integrated development environments (IDEs), plugins, and agentic interactions.
monitor embedded AI usage by users, websites and extensions
Related framework references (5)
Astrix Security claims it detects and responds to threats such as compromised credentials and out-of-scope agent actions.
Detect and respond to threats such as compromised credentials and out-of-scope agent actions.
Related framework references (5)