About Permiso
Permiso's Universal Identity Graph connects every identity to the credentials it owns, the machines it creates, the agents it runs and every action it takes in real time. Permiso states that the chain stays unbroken from a human SSO login to a Lambda function to an MCP tool call.[3]
Permiso states that modern threat actors are compromising identity providers and moving across environments through cloud service providers, SaaS apps and CI/CD pipelines. Permiso's library of integrations tracks activity across cloud environments. Permiso identifies the gap in agent security as what happens after authentication, where IdPs go blind to every tool call, MCP invocation and sub-agent spawn.[3]
Permiso Discover gives enterprises an inventory of human, non-human and AI identities wherever they live. Permiso Protect continuously evaluates the risk posture of identities based on usage patterns, exposure and entitlements. Permiso Protect surfaces and quantifies high-risk identities, stale access and toxic combinations across human, vendor, AI and non-human accounts. Permiso's Universal Identity Graph inventories identities in all cloud environments and assesses exposure risk. Permiso describes removing zombie accounts and identities, including vendors, that are no longer active in an environment. Permiso shows where MFA is not used and monitors the age and rotation of password and access keys. Permiso provides visualizations of different ways threat actors could compromise identities to move laterally and gain access to sensitive data.[1][3][4][6]
Permiso provides identity threat detection and response for human and non-human identities across environments. Permiso's detection engine has 1500+ detection rules engineered by its P0 Labs team, which actively researches and monitors attacks in the wild. Permiso detects access and behavioral anomalies in environments such as AWS, Azure, GCP, GitHub, Slack and Salesforce. Permiso monitors activity and correlates user behavior to produce identity security alerts across AWS, Azure, Okta and M365. Permiso tracks the entire lifecycle of credentials, their privileges and posture, and all of the sessions in which those credentials were used and by what identity. Permiso's watchlist capability identifies and mitigates threats from employees, contractors and other trusted individuals who may pose security risks to the organization. Permiso monitors human and non-human identities for access and behavior anomalies in AWS, GitHub, Jira, Salesforce and Snowflake.[3][4][7][8]
Permiso defines non-human identities (NHIs) as any digital identity that authenticates and operates without direct human control. Permiso inventories human and non-human identities, including keys, tokens, service accounts, roles and groups. Permiso identifies orphaned, stale and overly permissive NHIs across cloud environments. Permiso provides a centralized dashboard to monitor NHIs. Permiso monitors the behavior of non-human identities and the humans using them to detect suspicious and malicious activity.[9][10]
Permiso ties every run, event and tool call to the identity behind it in real time across cloud, infrastructure, SaaS and on-prem environments. Permiso launched AI agent runtime security, bringing Identity Runtime Attribution to every agent in production. The Permiso platform gives security teams visibility into agent runs, tool calls and MCP invocations across the full agent lifecycle. Permiso finds agents in Lambda functions, containers and VMs, in the code repositories where they are defined, and behind the model endpoints where they execute. Permiso ties every agent action to a specific human, non-human or AI identity in real time. Permiso visualizes agent actions through the Universal Identity Graph and preserves them as a complete audit trail. SandyClaw runs every skill in a controlled sandbox before it touches the environment, recording every action, every tool call and every downstream request. Permiso's agent lifecycle covers discovering agents, attributing identity at runtime, detecting anomalies, sandboxing skills and enforcing controls, including kill switches. Autodesk is among the first enterprises to deploy the agent capabilities, securing AI agents across its products, workforce and cloud infrastructure.[1][3][4][11]
Permiso detects suspicious activity in Google Cloud Platform across Compute, Storage, App Engine, Secrets Manager, Logging, Network, SSO and IAM. Permiso monitors when GitHub repositories have been transferred outside the organization or updated to public. Permiso provides threat detection for 1Password's password management system. Permiso inventories users in Open AI and monitors their posture and corresponding risk. Permiso monitors only the control plane and is not exposed to any underlying or transaction data. Permiso encrypts control plane data at rest and in transit. ACV Auctions is a Permiso customer. Nutanix is a Permiso customer. Modern Health is a Permiso customer. Permiso states its mission as simplifying cloud security for technology and security professionals.[3][12][13]
Permiso on the record