Salt Security

O que é Salt Security?

Salt Security provides an API security platform extended to agentic AI environments. It discovers APIs, AI agents, MCP servers, models, and their relationships, then helps security teams assess exposure and posture and monitor activity across the path from agent to downstream API and data. Its purpose is to make the assets and actions involved in API and agentic workflows visible and support governance and threat response. The platform is designed to work with existing security and development tools, including gateways, WAFs, SIEMs, and Jira workflows.

O que Salt Security pode fazer?

API and agent discovery

Builds inventories of APIs across internal, external, shadow, and third-party environments, and discovers agents, MCP servers, and models from code through runtime.

Agentic Security Graph

Maps relationships among LLMs, MCP servers, agents, and the APIs they invoke, helping teams trace which systems and data an agent can reach.

Posture and configuration assessment

Identifies exposed MCP servers, over-permissioned agents, API misconfigurations, and configuration drift for security teams to review.

Runtime threat detection

Monitors API and agent activity for behavioral and logic-based threats, including unauthorized tool usage, mass data pulls, and API abuse.

AI interaction monitoring

Inspects production model interactions for prompt injection, unsanctioned model access, and sensitive data moving into or out of LLM calls.

Response and workflow integrations

Supports real-time blocking of detected agentic attacks and connects findings with existing gateways, WAFs, SIEMs, and Jira workflows.

Casos de uso

“Find unmanaged AI infrastructure”

Security teams can locate shadow agents and MCP servers, including externally exposed servers, and assess the systems and APIs those assets can reach.

“Review agent permissions and governance”

Use the mapped agent-to-API relationships to identify permissions that exceed intended business boundaries and maintain traceable activity for audit work.

“Investigate suspicious API behavior”

SOC and application security teams can examine behavioral activity such as refund enumeration or mass data access in context, then use detection findings to support response.

“Reduce API exposure during development”

Application teams can review discovered APIs, sensitive-data exposure, and posture findings to prioritize remediation as services change.

“Monitor sensitive data in AI workflows”

Teams can inspect model interactions and API traffic to identify sensitive information moving through agent workflows and investigate potential exposure.

Perguntas frequentes

What does Salt Security protect?

Salt addresses API security and agentic AI security, including APIs, AI agents, MCP servers, models, and activity across their interactions.

How does Salt map agent activity?

Its Agentic Security Graph connects agents, LLMs, MCP servers, and the APIs they invoke, providing context about the systems and data involved.

Can Salt work with existing security tools?

The platform is described as integrating with existing gateways, WAFs, cloud and edge tools, SIEMs, and Jira workflows.

Does Salt support blocking threats?

Salt describes automated, real-time blocking for detected agentic attacks, alongside monitoring and detection of API and agent behavior.