# AI Agent Access Control > AI Agent Access Control maps the identity, default permissions and boundaries of AI agent platforms in AWS, Azure and Google Cloud, with a 30-check checklist. Every page has a markdown twin at the same address plus index.md (or send Accept: text/markdown). Every fact carries its source and the date it was read (10 Nov 2026). Datasets: https://aiagentaccesscontrol.com/datasets/ (CSV and JSON, CC BY 4.0). ## Home - [AI agent access control: what an agent can reach in your cloud, and where to draw the line](https://aiagentaccesscontrol.com/index.md): AI Agent Access Control maps the identity, default permissions and boundaries of AI agent platforms in AWS, Azure and Google Cloud, with a 30-check checklist. ## Method - [Seven criteria, fixed weights, and coverage of identity, network and AI access carrying the most](https://aiagentaccesscontrol.com/method/index.md): How AI Agent Access Control scores: the seven criteria, the weights fixed before scoring, the blind scoring, the rules and the read dates. ## Directory - [Tools scored on AI access coverage](https://aiagentaccesscontrol.com/tools/index.md): The 14 platforms AI Agent Access Control scores, in rank order, each with its own page. - [Picks by situation](https://aiagentaccesscontrol.com/situations/index.md): The five buyer situations AI Agent Access Control re-weights its rubric for, with the top three in each. - [Guides](https://aiagentaccesscontrol.com/guides/index.md): Guides on putting an AI agent into production in AWS, Azure and Google Cloud, written from the same provider pages as the map. Each guide lists its sources with the date it was read and ends with the checklist items it closes. - [Glossary](https://aiagentaccesscontrol.com/glossary/index.md): Plain definitions of the terms in the map and checklist, from execution role and agent identity to permissions boundary, Principal Access Boundary and data event. Each entry has a sourced example from a cloud provider page read on 10 Nov 2026. - [News](https://aiagentaccesscontrol.com/news/index.md): Dated news that changes the map or the ranking: renames, retirements, preview launches and acquisitions, such as Bedrock Agents becoming Agents Classic and Apono joining 1Password in June 2026. Each item names its source and the date the page was read. - [Agent platforms in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/agent-platforms/index.md): 8 agent platforms and self-hosted patterns, each with the identity it runs as and the boundaries that apply. ## Platform - [Act Security for AI agent access: 1st of 14 at 5.3](https://aiagentaccesscontrol.com/tools/act/index.md): Act Security is 1st of 14 on AI access coverage at 5.3 of 10. It scores 6 on identity, network and AI access together, the criterion weighted 35 here, and - [Sonrai Security for AI agent access: 2nd of 14 at 4.8](https://aiagentaccesscontrol.com/tools/sonrai/index.md): Sonrai Security is 2nd of 14 on AI access coverage at 4.8 of 10. It scores 4 on identity, network and AI access together, the criterion weighted 35 here, a - [Wiz for AI agent access: 3rd of 14 at 4.2](https://aiagentaccesscontrol.com/tools/wiz/index.md): Wiz is 3rd of 14 on AI access coverage at 4.2 of 10. It scores 6 on identity, network and AI access together, the criterion weighted 35 here, and 3 on remo - [Google Cloud IAM recommender for AI agent access: 4th of 14 at 4.0](https://aiagentaccesscontrol.com/tools/google-iam-recommender/index.md): Google Cloud IAM recommender is 4th of 14 on AI access coverage at 4.0 of 10. It scores 4 on identity, network and AI access together, the criterion weight - [Apono for AI agent access: joint 5th of 14 at 3.9](https://aiagentaccesscontrol.com/tools/apono/index.md): Apono is joint 5th of 14 on AI access coverage at 3.9 of 10. It scores 4 on identity, network and AI access together, the criterion weighted 35 here, and 5 - [Tenable Cloud Security for AI agent access: joint 5th of 14 at 3.9](https://aiagentaccesscontrol.com/tools/tenable/index.md): Tenable Cloud Security is joint 5th of 14 on AI access coverage at 3.9 of 10. It scores 5 on identity, network and AI access together, the criterion weight - [Britive for AI agent access: 7th of 14 at 3.6](https://aiagentaccesscontrol.com/tools/britive/index.md): Britive is 7th of 14 on AI access coverage at 3.6 of 10. It scores 4 on identity, network and AI access together, the criterion weighted 35 here, and 5 on - [CrowdStrike for AI agent access: joint 8th of 14 at 3.3](https://aiagentaccesscontrol.com/tools/crowdstrike/index.md): CrowdStrike is joint 8th of 14 on AI access coverage at 3.3 of 10. It scores 4 on identity, network and AI access together, the criterion weighted 35 here, - [P0 Security for AI agent access: joint 8th of 14 at 3.3](https://aiagentaccesscontrol.com/tools/p0/index.md): P0 Security is joint 8th of 14 on AI access coverage at 3.3 of 10. It scores 4 on identity, network and AI access together, the criterion weighted 35 here, - [Prisma Cloud for AI agent access: joint 8th of 14 at 3.3](https://aiagentaccesscontrol.com/tools/prisma/index.md): Prisma Cloud is joint 8th of 14 on AI access coverage at 3.3 of 10. It scores 5 on identity, network and AI access together, the criterion weighted 35 here - [Entitle for AI agent access: joint 11th of 14 at 3.2](https://aiagentaccesscontrol.com/tools/entitle/index.md): Entitle is joint 11th of 14 on AI access coverage at 3.2 of 10. It scores 4 on identity, network and AI access together, the criterion weighted 35 here, an - [Sysdig for AI agent access: joint 11th of 14 at 3.2](https://aiagentaccesscontrol.com/tools/sysdig/index.md): Sysdig is joint 11th of 14 on AI access coverage at 3.2 of 10. It scores 4 on identity, network and AI access together, the criterion weighted 35 here, and - [Microsoft for AI agent access: 13th of 14 at 3.1](https://aiagentaccesscontrol.com/tools/microsoft/index.md): Microsoft is 13th of 14 on AI access coverage at 3.1 of 10. It scores 6 on identity, network and AI access together, the criterion weighted 35 here, and 2 - [AWS IAM Access Analyzer for AI agent access: 14th of 14 at 2.7](https://aiagentaccesscontrol.com/tools/aws-iam-access-analyzer/index.md): AWS IAM Access Analyzer is 14th of 14 on AI access coverage at 2.7 of 10. It scores 2 on identity, network and AI access together, the criterion weighted 3 ## Situation - [For a company putting AI agents into production this year, Act is first at 5.3 and Sonrai Security is second at 4.8](https://aiagentaccesscontrol.com/situations/ai-agents/index.md): For a company putting AI agents into production this year, the top three on this situation's weights are Act Security (5.3), Sonrai Security (4.8) and Wiz - [For an AWS estate with hundreds of accounts, Act is first at 5.3 and Sonrai Security is second at 5.1](https://aiagentaccesscontrol.com/situations/aws-scale/index.md): For an AWS-heavy organization with hundreds of accounts and years of accumulated access, the top three on this situation's weights are Act Security (5.3), - [For an agent estate across AWS, Azure and Google Cloud, Act is first at 5.3 and Sonrai Security is second at 5.0](https://aiagentaccesscontrol.com/situations/multi-cloud/index.md): For a multi-cloud enterprise with separate teams per cloud, the top three on this situation's weights are Act Security (5.3), Sonrai Security (5.0) and Goo - [For a regulated company that must prove its agent boundaries, Act is first at 4.9 and Sonrai Security is a close second at 4.8](https://aiagentaccesscontrol.com/situations/regulated/index.md): For a regulated company that must prove access boundaries for PCI DSS, HIPAA or NIST 800-53, the top three on this situation's weights are Act Security (4. - [For a small team that cannot clear its findings, Act is first at 5.2 and Sonrai Security is second at 5.1](https://aiagentaccesscontrol.com/situations/small-team/index.md): For a small cloud security team buried under posture findings it cannot clear, the top three on this situation's weights are Act Security (5.2), Sonrai Sec ## Guide - [The identity an AI agent runs as in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/guides/the-identity-an-agent-runs-as/index.md): An agent has no access of its own. Every call it makes is judged against the identity it runs as: an IAM role on AWS, a service account or agent identity o - [Why an AI agent's permissions are not its user's permissions](https://aiagentaccesscontrol.com/guides/why-an-agents-permissions-are-not-its-users/index.md): A user who talks to an agent does not lend it their access. The agent calls tools with its own identity unless the platform exchanges tokens to act for the - [Network reachability for AI agents on AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/guides/network-reachability-for-agents/index.md): What an agent can reach over the network depends on where it runs and what you attach. This guide covers private endpoints, VPC attachment and egress filte - [Logging what an AI agent did: audit, data and model logs](https://aiagentaccesscontrol.com/guides/logging-what-an-agent-did/index.md): After an incident the question is which agent did what, to which data. This guide covers the three log layers on AWS, Google Cloud and Azure (control plane - [AI agent permissions: how to cut an agent's access](https://aiagentaccesscontrol.com/guides/least-privilege-for-ai-agents/index.md): AI agent permissions start too wide when a starter role or broad managed policy gets the agent working fast. This guide shows how to write the policy from - [Amazon Bedrock agent permissions: roles, trust and what the console creates](https://aiagentaccesscontrol.com/guides/bedrock-agent-permissions/index.md): Bedrock agents run as IAM roles that the service assumes. This guide covers Agents Classic and AgentCore: the trust policy and its condition keys, the docu - [Vertex AI agent service account: Agent Runtime identities in Google Cloud](https://aiagentaccesscontrol.com/guides/vertex-ai-agent-service-account/index.md): Agent Runtime, formerly Vertex AI Agent Engine, lets an agent run as a per-agent identity, a Google-managed service agent or a custom service account. This - [What can an AI agent access in AWS: the four layers to read](https://aiagentaccesscontrol.com/guides/what-can-an-ai-agent-access-in-aws/index.md): To answer what an agent can reach in an AWS account, read four layers: the role it assumes, the tools and Lambda functions behind it, the resource-based po - [Approving an AI agent for production: running the review meeting](https://aiagentaccesscontrol.com/guides/approving-an-agent-for-production/index.md): A production review for an AI agent works when it follows a fixed order and ends with recorded decisions. This guide gives the pre-read, the agenda built o - [AI agent permissions in Azure: Foundry Agent Service, identities and RBAC](https://aiagentaccesscontrol.com/guides/ai-agent-permissions-in-azure/index.md): Microsoft Foundry Agent Service gives each agent a Microsoft Entra agent identity that starts without roles. This guide covers agent identities, project an - [AI agent authorization in AWS, Azure and Google Cloud](https://aiagentaccesscontrol.com/ai-agent-authorization/index.md): AI agent authorization in a cloud account is one decision made on up to three inputs: the identity the agent runs as, the policies attached to the resource - [AI agent guardrails: SCPs, Azure Policy and org policy](https://aiagentaccesscontrol.com/ai-agent-guardrails/index.md): AI agent guardrails for cloud access are the limits set above the agent's own role: AWS SCPs, RCPs and permissions boundaries, Azure Policy and deny assign - [AI access control in cloud accounts: roles, policy, network](https://aiagentaccesscontrol.com/ai-access-control/index.md): AI access control for AI workloads in a cloud account has four layers: the identity the workload runs as, the policies that grant it access, the ceilings a - [Agent authorization patterns: five ways to scope an agent](https://aiagentaccesscontrol.com/agent-authorization-patterns/index.md): Five authorization patterns scope an AI agent in a cloud account: a dedicated role per agent, a role per tool, session policies that scope down credentials - [Amazon Q Business default permissions: roles it creates](https://aiagentaccesscontrol.com/default-permissions/amazon-q-business/index.md): Amazon Q Business runs as separate IAM service roles for the application, web experience, data sources and plugins, which the console creates if you let it - [Copilot Studio default permissions: tool authentication](https://aiagentaccesscontrol.com/default-permissions/copilot-studio/index.md): In Microsoft Copilot Studio an agent's reach depends on how each tool authenticates: agent author authentication, which Microsoft recommends for low-risk o - [Azure OpenAI Assistants default permissions and access](https://aiagentaccesscontrol.com/default-permissions/azure-openai-assistants/index.md): Azure OpenAI Assistants (Preview) is documented by Microsoft as retired in favor of Foundry Agent Service. The quickstart authenticates with an API key; En - [Conversational Agents (Dialogflow CX) default permissions](https://aiagentaccesscontrol.com/default-permissions/conversational-agents/index.md): Conversational Agents (Dialogflow CX) calls tools and webhooks as the Dialogflow CX Service Agent, or as a service account you name. Google documents that ## Glossary - [AI agent: meaning and an example](https://aiagentaccesscontrol.com/glossary/ai-agent/index.md): A program that uses a language model to decide which actions to take, then carries them out by calling tools, APIs and data sources on its own. In a cloud - [AI workload: meaning and an example](https://aiagentaccesscontrol.com/glossary/ai-workload/index.md): Any program running in a cloud account whose job involves a model: an agent, a retrieval service, a batch scoring job or an inference endpoint. It gets an - [Execution role: meaning and an example](https://aiagentaccesscontrol.com/glossary/execution-role/index.md): An IAM role that a function or agent runtime assumes when it runs. The role's policies decide which AWS APIs the code can call, and the service assumes the - [Service role: meaning and an example](https://aiagentaccesscontrol.com/glossary/service-role/index.md): An IAM role that an AWS service assumes to act in your account on your behalf. The role has a trust policy naming the service and permission policies listi - [Service account: meaning and an example](https://aiagentaccesscontrol.com/glossary/service-account/index.md): A non-human identity in Google Cloud that a workload runs as. You grant it IAM roles, attach it to a resource such as a Cloud Run service, and the workload - [Service agent: meaning and an example](https://aiagentaccesscontrol.com/glossary/service-agent/index.md): A Google-managed service account that a Google Cloud service uses to act in your project, such as the account Agent Runtime uses by default. Google creates - [Managed identity: meaning and an example](https://aiagentaccesscontrol.com/glossary/managed-identity/index.md): An Azure identity assigned to a compute resource so that the resource can get Microsoft Entra tokens without any stored secret. It is either system-assigne - [Workload identity: meaning and an example](https://aiagentaccesscontrol.com/glossary/workload-identity/index.md): An identity held by software instead of a person. Cloud providers issue them to applications, containers and functions so that code can authenticate to oth - [Workload identity federation: meaning and an example](https://aiagentaccesscontrol.com/glossary/workload-identity-federation/index.md): A way for software running outside a cloud, or in another system, to trade a token from its own identity provider for short-lived cloud credentials. The cl - [EKS Pod Identity: meaning and an example](https://aiagentaccesscontrol.com/glossary/eks-pod-identity/index.md): An Amazon EKS feature that maps an IAM role to a Kubernetes service account in a namespace, so the pods using that service account receive the role's tempo - [IAM roles for service accounts: meaning and an example](https://aiagentaccesscontrol.com/glossary/iam-roles-for-service-accounts/index.md): The earlier Amazon EKS method, often shortened to IRSA, for giving a pod an IAM role. A Kubernetes service account is linked to a role, and the pod's conta - [Action group: meaning and an example](https://aiagentaccesscontrol.com/glossary/action-group/index.md): In Amazon Bedrock Agents, the set of actions an agent may perform, defined by an API schema or function details and usually backed by a Lambda function. Ea - [Agent tool: meaning and an example](https://aiagentaccesscontrol.com/glossary/agent-tool/index.md): A function, API or service that an agent can call to act or to fetch data. The tool runs under some identity, which may be the agent's, the user's or its o - [Agent runtime: meaning and an example](https://aiagentaccesscontrol.com/glossary/agent-runtime/index.md): A managed cloud service that hosts agent code, runs it for each session and gives it an identity and a network position. Google Cloud calls its service Age - [On-behalf-of flow: meaning and an example](https://aiagentaccesscontrol.com/glossary/on-behalf-of-flow/index.md): An authorization pattern where a service exchanges the token of the signed-in user for a new token aimed at a downstream resource, so the downstream call i - [Identity propagation: meaning and an example](https://aiagentaccesscontrol.com/glossary/identity-propagation/index.md): Carrying who is asking, and who is acting, through each hop between a user, an agent and the services it calls. Propagation lets a downstream service apply - [Model invocation logging: meaning and an example](https://aiagentaccesscontrol.com/glossary/model-invocation-logging/index.md): An Amazon Bedrock setting that records the full request data, response data and metadata of model calls in an account and Region. It delivers to CloudWatch - [CloudTrail data events: meaning and an example](https://aiagentaccesscontrol.com/glossary/cloudtrail-data-events/index.md): AWS CloudTrail records of operations on or inside a resource, such as reading an S3 object or invoking a function, as opposed to management events that rec - [Cloud Audit Logs: meaning and an example](https://aiagentaccesscontrol.com/glossary/cloud-audit-logs/index.md): Google Cloud's audit trail of who did what, where and when. It has several log types, including Admin Activity logs for configuration changes and Data Acce - [Data access audit logs: meaning and an example](https://aiagentaccesscontrol.com/glossary/data-access-audit-logs/index.md): The Google Cloud audit logs that record reads and writes of user-provided data. They can be large, so they are disabled by default for most services and mu - [VPC endpoint: meaning and an example](https://aiagentaccesscontrol.com/glossary/vpc-endpoint/index.md): A private connection from an AWS VPC to an AWS service or another service, without a route through an internet gateway. An endpoint can carry its own polic - [AWS PrivateLink: meaning and an example](https://aiagentaccesscontrol.com/glossary/aws-privatelink/index.md): The AWS technology behind interface VPC endpoints. It gives a private connection between a VPC and a service, so traffic reaches the service as if it were - [Private Service Connect: meaning and an example](https://aiagentaccesscontrol.com/glossary/private-service-connect/index.md): A Google Cloud networking capability that lets consumers reach managed services privately from their own VPC networks using internal IP addresses. Producer - [Private endpoint: meaning and an example](https://aiagentaccesscontrol.com/glossary/private-endpoint/index.md): An Azure network interface with a private IP address from your virtual network that connects privately to a service on Azure Private Link. It brings the se - [Egress control: meaning and an example](https://aiagentaccesscontrol.com/glossary/egress-control/index.md): Limiting where outbound traffic from a workload may go, by destination, protocol or domain. For an agent, egress rules decide whether a hijacked prompt can - [Knowledge base: meaning and an example](https://aiagentaccesscontrol.com/glossary/knowledge-base/index.md): In Amazon Bedrock, a store of your documents that an agent can query for extra context when it builds an answer. Access to it is a permission on the agent' - [Retrieval augmented generation: meaning and an example](https://aiagentaccesscontrol.com/glossary/retrieval-augmented-generation/index.md): A pattern where a system looks up relevant documents first and gives them to a language model along with the question, so the answer draws on those documen - [Confused deputy: meaning and an example](https://aiagentaccesscontrol.com/glossary/confused-deputy/index.md): A security problem where something without permission to do an action tricks a more-privileged party into doing it. A cloud service acting on your behalf i - [Resource-based policy: meaning and an example](https://aiagentaccesscontrol.com/glossary/resource-based-policy/index.md): A permission policy attached to the resource being accessed, such as a Lambda function, an S3 bucket or a KMS key, naming who may use it. It works alongsid ## News - [AWS Nitro Sandboxes give pods, tasks and agents their own hardware-isolated VM](https://aiagentaccesscontrol.com/news/nitro-sandboxes-eks-ecs-ec2-agents/index.md): AWS announced AWS Nitro Sandboxes on 9 November 2026: lightweight VMs managed by the Nitro Hypervisor for an EKS pod, ECS task or VM, including AI agents r - [Amazon Quick adds smart approvals for actions that agents take](https://aiagentaccesscontrol.com/news/quick-smart-approvals-agentic-actions/index.md): AWS announced on 28 October 2026 that Amazon Quick supports smart approvals. A separate classifier model reviews each eligible agent action, approves routi - [WorkSpaces for AI agents adds OAuth sessions inside existing network boundaries](https://aiagentaccesscontrol.com/news/workspaces-for-ai-agents-oauth-sessions/index.md): AWS said on 20 October 2026 that WorkSpaces for AI agents now handles authentication and desktop connections in its managed MCP service. Sessions use exist - [Act Security and HUB International build a no-credentials coding agent environment](https://aiagentaccesscontrol.com/news/act-hub-international-agentcore-coding-agents/index.md): Act Security published a blog post on 19 October 2026 about HUB International moving Claude Code into cloud-hosted sandboxes on Amazon Bedrock AgentCore in - [Act Security argues AI governance should limit what agents can reach](https://aiagentaccesscontrol.com/news/act-real-ai-risk-what-agents-can-reach/index.md): Act Security published a news piece on 18 October 2026 titled The Real AI Risk Isn't How Agents Think. It's What They Can Reach. Act argues that agent reas - [Amazon Bedrock Managed Agents powered by OpenAI enters preview](https://aiagentaccesscontrol.com/news/bedrock-managed-agents-openai-preview/index.md): AWS announced on 29 September 2026 a preview of Amazon Bedrock Managed Agents, built jointly with OpenAI. AWS says each agent runs with its own IAM role, s - [Security Command Center flags reasoning engines that can modify IAM policies](https://aiagentaccesscontrol.com/news/scc-risk-engine-reasoning-engines-modify-iam/index.md): Google Cloud said on 28 September 2026 that the Security Command Center Risk Engine now detects and reports reasoning engines that can modify IAM policies - [IAM outbound identity federation supports VPC endpoints for OIDC discovery](https://aiagentaccesscontrol.com/news/iam-outbound-federation-vpc-endpoints-oidc/index.md): AWS announced on 25 September 2026 that IAM outbound identity federation supports interface VPC endpoints for OIDC discovery and JWKS key endpoints through - [New AgentCore Runtime keeps hardware-enforced session isolation](https://aiagentaccesscontrol.com/news/agentcore-runtime-v2-session-isolation/index.md): AWS announced on 18 September 2026 the next generation of AgentCore Runtime, the serverless microVM compute in Amazon Bedrock AgentCore. It keeps hardware- - [Policy Troubleshooter can now explain access denials for agents](https://aiagentaccesscontrol.com/news/policy-troubleshooter-agent-access-denials/index.md): Google Cloud announced on 15 September 2026 that Policy Troubleshooter supports troubleshooting access for agent identities. You can check IAM allow polici - [Google agent platform shell and computer use sandboxes reach GA with VPC-SC](https://aiagentaccesscontrol.com/news/agent-platform-sandboxes-vpc-sc-psc-ga/index.md): Google Cloud made Computer Use and Shell sandboxes in Gemini Enterprise Agent Platform generally available on 9 September 2026. The release adds VPC Servic - [Google IAM unified access policies for agents reach general availability](https://aiagentaccesscontrol.com/news/agent-platform-iam-access-policies-ga/index.md): Google Cloud made IAM Unified Access Policies for Gemini Enterprise Agent Platform generally available on 31 August 2026. Agent Gateway uses Identity-Aware - [AgentCore Memory adds fine-grained access control through Cedar policies](https://aiagentaccesscontrol.com/news/agentcore-memory-fine-grained-access-control/index.md): AWS announced on 28 August 2026 that Amazon Bedrock AgentCore Memory supports fine-grained access control. Placing a Memory resource behind an AgentCore Ga - [Amazon Quick custom permissions can deny new AI capabilities by default](https://aiagentaccesscontrol.com/news/quick-custom-permissions-deny-by-default/index.md): AWS announced on 19 August 2026 that Amazon Quick custom permissions now include deny by default. Administrators can restrict an AI capability category in ## Faq - [30 questions about ai agent access in the cloud](https://aiagentaccesscontrol.com/faq/index.md): Thirty answers for the security lead and the engineer: what identity a Bedrock, Agent Runtime or Foundry agent runs as, what it gets by default, how to log ## Dataset - [The Agent Access Map: the data, downloadable](https://aiagentaccesscontrol.com/datasets/index.md): Download The Agent Access Map and the scores as CSV and JSON with a cite-this line. Edition 10 Nov 2026. ## About - [About](https://aiagentaccesscontrol.com/about/index.md): AI Agent Access Control is a reference on what an AI agent can reach once it runs in AWS, Azure or Google Cloud. It has the Agent Access Map, the 30-check - [Editorial policy](https://aiagentaccesscontrol.com/editorial-policy/index.md): Scores come from a published rubric with fixed weights. Every claim about a product is attributed to the product's own pages or to a named report and carri - [Corrections](https://aiagentaccesscontrol.com/corrections/index.md): Send a correction to the editor with the page, the sentence and a source. A confirmed error is fixed, the old and new text are recorded here with the date, - [Contact](https://aiagentaccesscontrol.com/contact/index.md): Write to editor@aiagentaccesscontrol.com for corrections, evidence a vendor wants considered, or questions about a score or a map entry. Include the page a - [Privacy](https://aiagentaccesscontrol.com/privacy/index.md): AI Agent Access Control sets no cookies and runs no analytics script. The web server keeps ordinary access logs (address, page, time and browser string) fo - [Terms](https://aiagentaccesscontrol.com/terms/index.md): The content is a reference, as of the read dates shown. Cloud services change quickly, so check the provider's current documentation before you approve an ## Asset - [The Agent Access Map: what each agent platform runs as and can reach](https://aiagentaccesscontrol.com/map/index.md): The Agent Access Map lists eight agent setups across AWS, Google Cloud and Azure: five managed agent platforms and three self-hosted patterns. For each it - [The Agent Boundary Checklist: 30 checks before an agent goes to production](https://aiagentaccesscontrol.com/checklist/index.md): The Agent Boundary Checklist has 30 checks in seven sections: identity, permissions, data, network, tools, change control and logging. Each check says why ## Agent-platform - [Amazon Bedrock Agents (now Amazon Bedrock Agents Classic): identity, permissions and boundaries](https://aiagentaccesscontrol.com/agent-platforms/bedrock-agents/index.md): AWS documents that the agent runs as an IAM service role that Amazon Bedrock assumes. The documented service role policy covers invoking named foundation m - [Amazon Bedrock AgentCore (Runtime, Gateway, Identity): identity, permissions and boundaries](https://aiagentaccesscontrol.com/agent-platforms/bedrock-agentcore/index.md): Runtime runs the agent under an IAM execution role that AgentCore assumes. AWS documents that the console can create the execution role for you if the cons - [Agent Runtime on Gemini Enterprise Agent Platform (formerly Vertex AI Agent Engine, Agent Builder): identity, permissions and boundaries](https://aiagentaccesscontrol.com/agent-platforms/agent-runtime-gcp/index.md): Google documents three choices: a per-agent IAM agent identity (recommended, a principal tied to the individual agent); the Google-managed AI Platform Reas - [Gemini Enterprise (agents hosted on Agent Runtime): identity, permissions and boundaries](https://aiagentaccesscontrol.com/agent-platforms/gemini-enterprise/index.md): Gemini Enterprise registers an ADK agent that runs on Agent Runtime, so the agent runs under the Agent Runtime identity described in that entry. Permission - [Microsoft Foundry Agent Service (formerly Azure AI Foundry Agent Service): identity, permissions and boundaries](https://aiagentaccesscontrol.com/agent-platforms/foundry-agent-service/index.md): Each agent gets a Microsoft Entra agent identity, a special service principal created under an agent blueprint. Microsoft documents that a newly created ag - [Self-hosted agent on AWS Lambda or Amazon EKS: identity, permissions and boundaries](https://aiagentaccesscontrol.com/agent-platforms/self-hosted-aws/index.md): A Lambda function runs as its execution role. When you create a function in the Lambda console, AWS documents that it creates an execution role with minima - [Self-hosted agent on Google Cloud Run or GKE: identity, permissions and boundaries](https://aiagentaccesscontrol.com/agent-platforms/self-hosted-gcp/index.md): Cloud Run runs the service as a service account (the Compute Engine default service account if you do not choose one). Google documents that, depending on - [Self-hosted agent on Azure Container Apps or AKS: identity, permissions and boundaries](https://aiagentaccesscontrol.com/agent-platforms/self-hosted-azure/index.md): A container app can have a system-assigned managed identity (tied to the app) or user-assigned identities (standalone resources). Microsoft documents that ## Check - [Give each agent its own role or service account, not a shared one: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/own-identity-per-agent/index.md): Check 1 of 30: Give each agent its own role or service account, not a shared one. A shared identity means one agent's access is every agent's access, and l - [Use short-lived role credentials and no long-lived access keys for the agent: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/short-lived-credentials/index.md): Check 2 of 30: Use short-lived role credentials and no long-lived access keys for the agent. A key stored in code, a config file or a container image stays - [Limit who can pass, attach or impersonate the agent's role: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/restrict-who-can-attach-role/index.md): Check 3 of 30: Limit who can pass, attach or impersonate the agent's role. Anyone who can attach the agent's role to a new resource can borrow its access w - [Scope the role's trust to the specific agent, service account or workload: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/scope-trust-to-the-agent/index.md): Check 4 of 30: Scope the role's trust to the specific agent, service account or workload. A role that any workload in an account or cluster can assume is a - [Do not run the agent as a default or pre-granted account: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/no-default-accounts/index.md): Check 5 of 30: Do not run the agent as a default or pre-granted account. Default accounts are created for convenience and can carry broad roles that nobody - [Write the agent's permission policy from what its tools actually call: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/least-privilege-policy/index.md): Check 6 of 30: Write the agent's permission policy from what its tools actually call. Broad starter policies make every tool call possible, including ones - [Set an organization-level ceiling the agent's role cannot exceed: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/set-permission-ceiling/index.md): Check 7 of 30: Set an organization-level ceiling the agent's role cannot exceed. A role policy can be edited by whoever manages that account, so it is a we - [Explicitly deny the agent IAM changes and destructive actions: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/deny-iam-changes-and-destructive-actions/index.md): Check 8 of 30: Explicitly deny the agent IAM changes and destructive actions. An agent that can edit roles or delete data can widen its own access or cause - [Compare the agent's granted permissions with what it used and remove the rest: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/remove-unused-permissions/index.md): Check 9 of 30: Compare the agent's granted permissions with what it used and remove the rest. Permissions granted during a pilot tend to stay long after th - [Grant roles at the narrowest scope that works: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/narrowest-scope/index.md): Check 10 of 30: Grant roles at the narrowest scope that works. A role granted on a whole account, project or subscription covers every resource in it, incl - [Keep the encryption keys for agent data under your control and limit who can use them: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/control-the-keys/index.md): Check 11 of 30: Keep the encryption keys for agent data under your control and limit who can use them. If the agent's role can use a key, it can read whate - [Block public access on storage the agent can write to: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/block-public-exposure/index.md): Check 12 of 30: Block public access on storage the agent can write to. An agent that can change a bucket setting, or write to a bucket that already allows - [Put data stores the agent reads inside a perimeter that blocks copying data out: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/data-perimeter/index.md): Check 13 of 30: Put data stores the agent reads inside a perimeter that blocks copying data out. An agent allowed to read data can usually also send it som - [Limit data access with tags, attributes or conditions, not only resource lists: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/condition-based-data-access/index.md): Check 14 of 30: Limit data access with tags, attributes or conditions, not only resource lists. Resource lists go stale as new data stores appear. Conditio - [Run agent compute on private networks with no direct path to the internet: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/private-network-no-direct-internet/index.md): Check 15 of 30: Run agent compute on private networks with no direct path to the internet. Compute with a direct internet path can send data out or pull in - [Allow outbound traffic only to the destinations the agent needs: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/allowlist-egress/index.md): Check 16 of 30: Allow outbound traffic only to the destinations the agent needs. Open outbound access lets a hijacked agent reach any server to send data o - [Reach cloud APIs and data stores through private endpoints: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/private-endpoints-for-cloud-apis/index.md): Check 17 of 30: Reach cloud APIs and data stores through private endpoints. Calls to cloud services over public addresses can be made from anywhere that ho - [Keep the agent's code from reading node, instance or other workloads' credentials: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/credential-isolation/index.md): Check 18 of 30: Keep the agent's code from reading node, instance or other workloads' credentials. If agent code can read the host's metadata service, it c - [Give each tool function its own least-privilege role instead of the agent's role: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/tool-own-role/index.md): Check 19 of 30: Give each tool function its own least-privilege role instead of the agent's role. A tool that runs with the agent's full role can do everyt - [Restrict which principals can invoke each tool endpoint: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/restrict-tool-invokers/index.md): Check 20 of 30: Restrict which principals can invoke each tool endpoint. A tool endpoint that accepts calls from any caller can be used without the agent. - [Cap how much the agent can run and spend: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/cap-run-and-spend/index.md): Check 21 of 30: Cap how much the agent can run and spend. A looping or hijacked agent can start thousands of runs and a large bill before anyone notices. L - [Decide for each tool whether the action runs with the agent's authority or the signed-in user's: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/decide-user-or-agent-authority/index.md): Check 22 of 30: Decide for each tool whether the action runs with the agent's authority or the signed-in user's. If a tool runs with the agent's authority, - [Run agents in their own account, project or subscription: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/dedicated-account-project-subscription/index.md): Check 23 of 30: Run agents in their own account, project or subscription. Agents placed beside production workloads share their blast radius, quotas and ad - [Tag every agent resource with an owner and environment, and enforce the tags: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/tag-agent-resources/index.md): Check 24 of 30: Tag every agent resource with an owner and environment, and enforce the tags. Untagged agent resources cannot be traced to an owner when so - [Test permission and policy changes in a dry run before enforcing them: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/test-before-enforcing/index.md): Check 25 of 30: Test permission and policy changes in a dry run before enforcing them. A new deny or boundary can break an agent's tools in ways nobody saw - [Remove the agent's identity and access when the agent is retired: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/retire-identity-with-agent/index.md): Check 26 of 30: Remove the agent's identity and access when the agent is retired. Leftover identities from retired agents keep their permissions and nobody - [Turn on organization-wide audit logging of control-plane actions: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/control-plane-audit-logs/index.md): Check 27 of 30: Turn on organization-wide audit logging of control-plane actions. Without audit logs there is no record of which role or setting the agent - [Log reads and writes to the data the agent touches: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/log-data-access/index.md): Check 28 of 30: Log reads and writes to the data the agent touches. Control-plane logs show configuration changes, not which records the agent read. Data-l - [Make every logged action traceable to the specific agent identity: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/attribute-actions-to-the-agent/index.md): Check 29 of 30: Make every logged action traceable to the specific agent identity. If logs only show a shared role or a generic account, you cannot tell wh - [Alert on denied calls and on changes to the agent's role or policies: the check, and how in AWS, Google Cloud and Azure](https://aiagentaccesscontrol.com/checklist/alert-on-denied-and-role-changes/index.md): Check 30 of 30: Alert on denied calls and on changes to the agent's role or policies. A string of denied calls or an edit to the agent's role is often the ## Cloud - [What an AI agent can access in AWS](https://aiagentaccesscontrol.com/clouds/aws/index.md): In AWS, an AI agent runs as an IAM role, and can reach whatever that identity's permissions and the network allow. The map covers 3 ways to run one: Amazon - [What an AI agent can access in Google Cloud](https://aiagentaccesscontrol.com/clouds/gcp/index.md): In Google Cloud, an AI agent runs as a service account or a per-agent identity, and can reach whatever that identity's permissions and the network allow. T - [What an AI agent can access in Azure](https://aiagentaccesscontrol.com/clouds/azure/index.md): In Azure, an AI agent runs as a managed identity or an Entra agent identity, and can reach whatever that identity's permissions and the network allow. The