Collect Video Testimonials That Build Trust
Noma Security
AI agent security and governance platform for discovering assets, managing permissions, detecting risky runtime behavior, and testing applications with multi-turn attacks.
About Noma Security
Noma Security helps enterprises discover, govern, and protect AI agents across endpoint, cloud, and SaaS environments. Its platform combines AI Security Posture Management, Agent Access Control, AI Detection and Response, and AI Red Teaming. These capabilities connect information about an agent’s permissions and configuration with its actions during operation.
Noma evaluates runtime activity using context such as identity, session history, data access, and intended tasks. Depending on policy, responses can include blocking actions, masking sensitive information, redirecting an agent, requesting human involvement, or stopping activity.
The platform also examines agent supply chains and tests applications through multi-turn attack campaigns. Findings from these assessments can inform runtime policies as teams update models, instructions, and connected tools.
Key Features
- AI asset discovery: Identifies agents, models, MCP servers, skills, and toolsets across supported environments.
- Security posture assessment: Finds risky configurations, excessive permissions, exposed secrets, and problematic combinations of capabilities.
- Agent access control: Defines allowed actions using identity and permissions associated with agents and their owners.
- Contextual runtime detection: Evaluates actions against session context, data access, and agent intent.
- Flexible response policies: Supports blocking, masking, redirection, human escalation, and stopping agent activity.
- Multi-turn red teaming: Tests interactions that unfold over several steps and feeds findings into runtime security policies.
Pricing
Pricing not publicly specified.
Use Cases
- Discover AI agents created across cloud services and business applications.
- Review agent permissions before granting access to sensitive tools or data.
- Detect when an agent’s actions move beyond its intended task.
- Reassess AI applications after changes to models, instructions, or tool connections.
Pros & Cons
Pros:
- Connects configuration and permission context with runtime behavior.
- Covers endpoint, cloud, and SaaS agent environments.
- Supports multi-turn testing and several enforcement responses.
Integrations
AWS Bedrock, Azure AI Foundry, Databricks, LangChain, CrewAI, Microsoft Copilot Studio, Salesforce Agentforce, ServiceNow
FAQ
What is Noma Security used for?
Noma Security helps enterprises discover AI agents, assess their risks, control permissions, and enforce policies during operation.
Which environments does Noma cover?
Its platform describes coverage for endpoint agents, cloud applications, and agents built within supported SaaS platforms.
What does Noma AI-SPM assess?
It discovers AI assets and evaluates permissions, misconfigurations, supply chain components, and combinations of capabilities that can increase risk.
How does Noma AI-DR respond to risky activity?
Configured responses can include blocking actions, masking data, redirecting the agent, requesting human involvement, or stopping activity.
Can Noma restrict individual tool actions?
Yes. Agent Access Control supports permissions for specific actions, scoped by identity and enforced through supported integration points.
Does Noma support multi-turn red teaming?
Yes. It uses multi-turn campaigns to evaluate failures that may emerge later in an interaction and inform runtime policies.
Last edited
November 7, 2026 at 9:47 AM by Venkatraman Chandrasekaran
Similar to Noma Security
