Skip to content
IAM Policy Generator

The IAM policy simulator tests a policy without sending a request, and this workbench does not

AWS describes its IAM policy simulator as a tool that tests identity-based policies, permissions boundaries, SCPs and resource-based policies you supply, without sending a real request. This site's workbench is different: it reads and checks a policy against AWS reference data and does not evaluate it against an account.

Open the workbench ยท All guides

AWS's simulator returns allowed or denied for each action and resource

AWS documentation says the IAM policy simulator lets you test identity-based policies, IAM permissions boundaries, service control policies and resource-based policies that you provide. You simulate how AWS would evaluate a request against those policies without sending a real request to any AWS service. Each simulation returns a binary outcome for every action and resource you test, allowed or denied, and for an allow or an explicit deny it shows which policy produced the outcome.

The simulator has two modes. In Principal mode you choose an existing user, role or group, and the simulator evaluates the policies attached to it; you may add or exclude custom identity policies or a permissions boundary, which are simulated as if attached without changing the account. In Custom mode you write or paste one or more policies, which are simulated on their own and are not saved.

AWS lists where simulator results can differ from live behavior

AWS carries an Important note on the page: simulator results can differ from the live environment, and AWS recommends checking policies against the live environment after testing. The page names several differences. The simulator makes no real request, uses no real context key values from production and returns no service response. It evaluates SCPs, but it does not support resource control policies. Results can still differ for some advanced setups, such as VPC endpoint policies, role chaining or several resource-based policies on one resource.

For SCPs, the simulator reports whether an SCP allows or denies an action but, for security reasons, does not surface the matched statements as it does for other policy types. In the console you cannot set a value for a condition key that only an SCP references; with the AWS CLI or API you can.

Condition keys come from the simulator or from you

AWS says the simulator populates principal and organization context keys itself: aws:PrincipalAccount, aws:PrincipalId, aws:PrincipalType, aws:UserId and aws:UserName, and aws:PrincipalOrgID, aws:PrincipalOrgMasterAccountId and aws:PrincipalOrgPaths. You supply values for every other key your policies reference. If a value does not satisfy a condition, the action is denied.

The simulator also does not determine which services support a given global condition key for identity-based and resource-based policies. For example, it does not report that a service does not support aws:TagKeys. SCPs are evaluated with service-aware logic. The condition keys page covers the keys themselves.

Two AWS CLI commands run the same simulations from a terminal

AWS publishes two CLI commands for the simulator: aws iam simulate-principal-policy for policies attached to a principal and aws iam simulate-custom-policy for policies you pass in. The IAM actions iam:SimulatePrincipalPolicy and iam:SimulateCustomPolicy appear in the permissions boundary documentation as actions an administrator can allow or deny. Both commands call an account, which means they need AWS credentials and permission to run.

This workbench reads and checks a policy and does not evaluate it

The workbench on this site is not a simulator and does not evaluate a policy against an account. It runs in your browser, uploads nothing, and uses no credentials. It reads the JSON you paste, expands each wildcard against AWS's service authorization reference (457 services and 22,556 actions, fetched 16 Nov 2026), labels each action with its access level, and runs six checks on the policy text. The methodology page lists the six.

The two tools answer different questions. The AWS simulator answers whether a given action on a given resource would be allowed once the account's policies are applied. The workbench answers what a policy grants, how many actions a wildcard covers, and whether the text matches patterns AWS documents as risky.

QuestionAWS policy simulatorThis workbench
Does it send a request or call an accountNo request, but it runs in an AWS account and needs permissionsNo account call, no credentials, runs in the browser
What it returnsAllowed or denied for each action and resourceAction counts, access levels and findings on the text
SCPsEvaluates themReads and checks the text, including size and elements
RCPsNot supported, per AWSReads and checks the text, including size
Needs context key valuesYes, for keys it does not populateNo

Sources

  1. IAM policy testing with the IAM policy simulator, read 16 Nov 2026
  2. simulate-principal-policy, AWS CLI command reference, read 16 Nov 2026
  3. simulate-custom-policy, AWS CLI command reference, read 16 Nov 2026
  4. Permissions boundaries for IAM entities, read 16 Nov 2026
  5. AWS service reference information, read 16 Nov 2026

Last reviewed 16 Nov 2026.