These products address different jobs. Follow a control purpose to inspect its offerings, operating assumptions, and evidence. Placement groups related work; it does not indicate quality or rank.
Reviewed . Comparable edition scopes and published scoring criteria are needed before this market receives scores or momentum signals.
Each offering has one primary placement. Secondary tags show overlap within its documented scope. Filter by control purpose to include both primary and secondary coverage. Model supply-chain security is marked as adjacent.
24 offerings
No matching offerings. Clear filters or choose a different control purpose.
CONTROL PURPOSE 01
Agent discovery and security posture
Discover agents, owners, configurations and reachable resources before prioritizing exposure.
Where the boundary lies: An inventory or posture finding does not itself block an agent action. Connector scope determines what is visible.
Zenity combines agent inventory and posture assessment with exposure testing and runtime controls. This placement starts with discovering agents and assessing their reachable resources.
Noma describes discovery and posture assessment alongside agent access controls, red teaming and session detection. This profile begins with inventory and exposure context.
Prompt AI Agent Security is described as a discovery and governance control plane for agentic workflows. Its primary placement is visibility and policy context.
Lasso AI Security Posture Management maps agent and application exposure, including reachable data and available actions. Runtime protection is a complementary area of the platform.
Okta for AI Agents applies identity governance and least privilege to agent access. It makes the agent an explicit subject in the identity security model.
Astrix addresses agent and non-human identity discovery, access and lifecycle security. Its Cisco ownership does not make it equivalent to AI Defense validation.
Policy in AgentCore places an authorization decision at the gateway tool boundary. Cedar policies evaluate the principal, action, resource and relevant context before a tool call proceeds. This provides a concrete contrast with asking the language model to remember a business rule.
Bay describes endpoint discovery, session monitoring and action enforcement using session context. This profile emphasizes its allow, ask or deny decision at execution.
Zscaler AI Gateway places a control point in agent communications over MCP and A2A. Its scoped job is governing connections to tools, data and other agents.
Check Point AI Guardrails, identified as Lakera Guard in its privacy documentation, inspects submitted LLM inputs and outputs for threats and policy violations. The practical question is where an application sends content for inspection, what action follows a detection and how that content is retained.
Azure AI Content Safety provides APIs for application content checks. Prompt Shields addresses instructions that try to redirect a model, including attacks embedded in documents. The returned result is evidence for an application decision; a separate identity and permission system still controls access to records and tools.
Model Armor screens prompts and responses using configurable templates. Its inspect-only and blocking modes are useful for teaching the difference between a recorded policy violation and an interaction that was actually stopped. Integrations and content modalities have explicit boundaries that must be checked.
NeMo Guardrails is a programmable library for adding checks at several stages of an LLM application. Input, retrieval, dialog, execution and output rails expose different control points. That makes it an educational option for seeing exactly where a check belongs, while leaving configuration and operation with the application team.
Make it observable: The trace demonstrates which rail ran and whether the mock tool was called; a blocked answer alone is not counted as a blocked action.
Trace agent behavior, detect harmful activity and support intervention, investigation or recovery.
Where the boundary lies: Detection may occur after an action. Recovery depends on supported systems and preserved state; no placement implies every response function.
Falcon Guardian connects agent execution with Falcon telemetry for discovery, runtime detection and response. The endpoint execution chain is the primary scope here.
Rubrik Agent Cloud describes monitoring, governance and remediation across agent data, identities and actions. This profile connects behavioral evidence to intervention and recovery questions.
Proofpoint Agentic AI Security describes runtime observability and behavioral anomaly detection across multistep agent workflows. Its Acuvity lineage is ownership context.
AI Defense combines model and application validation with separately integrated runtime protection. Its management API exposes validation runs and results alongside application connections and policy events. This makes it useful for learning how a predeployment finding becomes a runtime control, without treating the two as the same test.
HiddenLayer AI Attack Simulation tests AI systems with adversarial cases, including prompt manipulation, data leakage and unsafe tool use. Its value to an evaluation is the reproducible failing interaction and resulting repair, not the existence of a report or an aggregate score.
Make it observable: The result identifies an observable unauthorized mock action, and a corrected policy prevents it on retest without breaking permitted use.
Promptfoo provides AI security testing and red teaming, including an open-source project. Its primary job is finding failures in a configured application.
Prisma AIRS spans several AI security modules. AI Model Security examines model artifacts and their supply chain before use, while runtime APIs inspect interactions after integration. A learner should keep an artifact finding separate from evidence that a deployed agent can access too much data.
Make it observable: The review preserves artifact identity, the finding and promotion decision; it makes no claim that the scan certifies runtime behavior.
An investigation agent reads a suspicious message, gathers context, and proposes a response. Trace which component checks the content, which identity authorizes each tool call, and which policy governs the proposed action.
In a lab, place conflicting instructions in synthetic evidence. Record whether the agent attempts an unauthorized action and whether the tool boundary prevents it. A model refusing the request and an authorization control denying it are different observations.
