Use the `aws_iam_policy` InSpec audit resource to test properties of a single managed AWS IAM Policy. Use `aws_iam_policies` to audit IAM policies in bulk.
A policy defines the permissions of an identity or resource within AWS. AWS evaluates these policies when a principal, such as a user, makes a request. Policy permissions, also called "policy statements" in AWS, determine if a request is authorized -- and allow or deny it accordingly.
Returns the default version of the policy document after decoding as a Ruby hash. This hash contains the policy statements and is useful for performing checks that cannot be expressed using higher-level matchers like `have_statement`.
For details regarding the contents of this structure, refer to the [AWS IAM Policy JSON Reference](https://docs.aws.amazon.com/IAM/latest/UserGuide/reference_policies.html). A set of examples is [also available](https://docs.aws.amazon.com/IAM/latest/UserGuide/access_policies_examples.html).
This InSpec audit resource has the following special matchers. For a full list of available matchers, please visit our [Universal Matchers page](https://www.inspec.io/docs/reference/matchers/).
Examines the list of statements contained in the policy and passes if at least one of the statements matches. This matcher does _not_ interpret the policy in a request authorization context, as AWS does when a request processed. Rather, `have_statement` examines the literal contents of the IAM policy, and reports on what is present (or absent, when used with `should_not`).
`have_statement` accepts the following criteria to search for matching statements. If any statement matches all the criteria, the test is successful.
* `Action` - Expresses the requested operation. Acceptable literal values are any AWS operation name, including the '*' wildcard character. `Action` may also use a list of AWS operation names.
* `Effect` - Expresses if the operation is permitted. Acceptable values are 'Deny' and 'Allow'.
* `Sid` - A user-provided string identifier for the statement.
* `Resource` - Expresses the operation's target. Acceptable values are ARNs, including the '*' wildcard. `Resource` may also use a list of ARN values.
Please note the following about the behavior of `have_statement`:
* `Action`, `Sid`, and `Resource` allow using a regular expression as the search critera instead of a string literal.
* it does not support wildcard expansion; to check for a wildcard value, check for it explicitly. For example, if the policy includes a statement with `"Action": "s3:*"` and the test checks for `Action: "s3:PutObject"`, the test _will not match_. You must write an additional test checking for the wildcard case.
* it supports searching list values. For example, if a statement contains a list of 3 resources, and a `have_statement` test specifes _one_ of those resources, it will match.
* `Action` and `Resource` allow using a list of string literals or regular expressions in a test, in which case _all_ must match on the _same_ statement for the test to match. Order is ignored.
* it does not support the `[Principal](https://docs.aws.amazon.com/IAM/latest/UserGuide/intro-structure.html#intro-structure-principal)` or `Conditional` key, or any of `NotAction`, `Not[Principal](https://docs.aws.amazon.com/IAM/latest/UserGuide/intro-structure.html#intro-structure-principal)`, or `NotResource`.
Your [Principal](https://docs.aws.amazon.com/IAM/latest/UserGuide/intro-structure.html#intro-structure-principal) will need the `iam:GetPolicy`, `iam:ListPolicy`, and `iam:ListEntitiesForPolicy` actions set to allow.
You can find detailed documentation at [Actions, Resources, and Condition Keys for Identity And Access Management](https://docs.aws.amazon.com/IAM/latest/UserGuide/list_identityandaccessmanagement.html).