implementing-aws-iam-permission-boundaries skill (Anthropic-Cybersecurity-Skills)
- Install
- SKILL.md (verbatim)
- Overview
- When to Use
- Prerequisites
- Core Concepts
- How Permission Boundaries Work
- Policy Evaluation Logic
- Key Use Cases
- Workflow
- Step 1: Define the Permission Boundary Policy
- Step 2: Create the Developer Delegation Policy
- Step 3: Attach the Boundary
- Step 4: Prevent Privilege Escalation
- Step 5: Deploy with Terraform
- Validation Checklist
- References
- Other files in this skill
- assets/template.md (verbatim)
- Boundary Policy Details
- Allowed Service Actions
- Denied Actions (Escalation Prevention)
- Testing Results
- Sign-Off
- references/api-reference.md (verbatim)
- Dependencies
- CLI Usage
- Functions
- getiamclient(profile, region)
- createpermissionboundary(client, policyname, allowedservices, allowedregions) -> dict
- attachboundarytorole(client, rolename, boundaryarn) -> dict
- auditroleswithoutboundary(client) -> list
- auditboundaryeffectiveness(client, rolename) -> dict
- generatereport(client) -> dict
- boto3 IAM Methods Used
- Output Schema
- references/standards.md (verbatim)
- AWS IAM Policy Types
- AWS Well-Architected Framework - Security Pillar
- SEC02 - Identity Management
- SEC03 - Permissions Management
- CIS AWS Foundations Benchmark v3.0
- NIST SP 800-53 Mapping
- references/workflows.md (verbatim)
- Boundary Policy Creation Workflow
- Developer Role Creation Workflow (with Boundary)
- Privilege Escalation Prevention Workflow
- Boundary Audit Workflow
What it does. Configures AWS IAM permission boundaries that cap the maximum permissions Part of mukul975/Anthropic-Cybersecurity-Skills (817 security skills) (mukul975/Anthropic-Cybersecurity-Skills).
| Upstream | mukul975/Anthropic-Cybersecurity-Skills |
| Skill file | skills/implementing-aws-iam-permission-boundaries/SKILL.md |
| License | Apache-2.0 (skill folder LICENSE) |
| Author | mukul975 |
| Fetched | 2026-09-10 |
Install
npx skills add mukul975/Anthropic-Cybersecurity-Skills --skill implementing-aws-iam-permission-boundaries, or copy the skill folder into~/.claude/skills/implementing-aws-iam-permission-boundaries/.- Raw file:
curl -sL https://raw.githubusercontent.com/mukul975/Anthropic-Cybersecurity-Skills/HEAD/skills/implementing-aws-iam-permission-boundaries/SKILL.md
SKILL.md (verbatim)
name: implementing-aws-iam-permission-boundaries
description: Configures AWS IAM permission boundaries that cap the maximum permissions
an identity-based policy can grant to a user or role, so effective permissions
are the intersection of the identity policy and the boundary even if the policy
grants AdministratorAccess. Use when letting security teams delegate IAM role and
policy creation to developers for self-service while enforcing least-privilege
ceilings and preventing privilege escalation.
domain: cybersecurity
subdomain: identity-access-management
tags:
- aws
- iam
- permission-boundaries
- least-privilege
- delegation
- cloud-security
version: '1.0'
author: mahipal
license: Apache-2.0
nist_csf:
- PR.AA-01
- PR.AA-02
- PR.AA-05
- PR.AA-06
mitre_attack:
- T1078
- T1110
- T1556
- T1098
- T1078.004
Implementing AWS IAM Permission Boundaries
Overview
IAM permission boundaries are an advanced AWS feature that sets the maximum permissions an identity-based policy can grant to an IAM entity (user or role). They enable centralized security teams to safely delegate IAM role and policy creation to application developers without risking privilege escalation. The effective permissions of an entity are the intersection of its identity-based policies and its permission boundary -- even if an identity policy grants AdministratorAccess, the permission boundary restricts it to only the allowed actions.
When to Use
- When deploying or configuring implementing aws iam permission boundaries capabilities in your environment
- When establishing security controls aligned to compliance requirements
- When building or improving security architecture for this domain
- When conducting security assessments that require this implementation
Prerequisites
- AWS account with IAM administrative access
- Understanding of AWS IAM policy language (JSON)
- AWS CLI v2 configured with appropriate credentials
- Terraform or CloudFormation for infrastructure-as-code deployment
Core Concepts
How Permission Boundaries Work
Identity-Based Policy Permission Boundary
(What the role CAN do) ∩ (What the role MAY do)
│ │
└──────────┬───────────────────┘
│
Effective Permissions
(Only actions in BOTH policies)
Policy Evaluation Logic
AWS evaluates permissions in this order:
- Explicit Deny in any policy - always wins
- Organizations SCP - sets org-wide maximum
- Permission Boundary - sets entity-level maximum
- Identity-Based Policy - grants actual permissions
- Resource-Based Policy - cross-account access (evaluated separately)
The entity can only perform an action if ALL applicable policy types allow it.
Key Use Cases
| Use Case | Description |
|---|---|
| Developer Delegation | Allow devs to create IAM roles without escalating beyond their boundary |
| Sandbox Isolation | Limit what roles can do in sandbox/dev accounts |
| Multi-Tenant Workloads | Ensure tenant-specific roles cannot access other tenants' resources |
| CI/CD Pipeline Roles | Restrict automation roles to specific services |
Workflow
Step 1: Define the Permission Boundary Policy
Create a managed policy that defines the maximum allowed permissions:
{
"Version": "2012-10-17",
"Statement": [
{
"Sid": "AllowedServices",
"Effect": "Allow",
"Action": [
"s3:*",
"dynamodb:*",
"lambda:*",
"logs:*",
"cloudwatch:*",
"sqs:*",
"sns:*",
"events:*",
"states:*",
"xray:*",
"ec2:Describe*",
"ec2:CreateTags",
"sts:AssumeRole",
"kms:Decrypt",
"kms:GenerateDataKey",
"kms:DescribeKey",
"secretsmanager:GetSecretValue"
],
"Resource": "*"
},
{
"Sid": "AllowIAMPassRole",
"Effect": "Allow",
"Action": "iam:PassRole",
"Resource": "arn:aws:iam::*:role/app-*",
"Condition": {
"StringEquals": {
"iam:PassedToService": [
"lambda.amazonaws.com",
"states.amazonaws.com"
]
}
}
},
{
"Sid": "DenyBoundaryDeletion",
"Effect": "Deny",
"Action": [
"iam:DeletePolicy",
"iam:DeletePolicyVersion",
"iam:CreatePolicyVersion"
],
"Resource": "arn:aws:iam::*:policy/DeveloperBoundary"
},
{
"Sid": "DenyBoundaryRemoval",
"Effect": "Deny",
"Action": [
"iam:DeleteUserPermissionsBoundary",
"iam:DeleteRolePermissionsBoundary"
],
"Resource": "*"
}
]
}
Step 2: Create the Developer Delegation Policy
Grant developers the ability to create IAM roles, but only with the boundary attached:
{
"Version": "2012-10-17",
"Statement": [
{
"Sid": "AllowCreateRoleWithBoundary",
"Effect": "Allow",
"Action": [
"iam:CreateRole",
"iam:AttachRolePolicy",
"iam:DetachRolePolicy",
"iam:PutRolePolicy",
"iam:DeleteRolePolicy"
],
"Resource": "arn:aws:iam::*:role/app-*",
"Condition": {
"StringEquals": {
"iam:PermissionsBoundary": "arn:aws:iam::*:policy/DeveloperBoundary"
}
}
},
{
"Sid": "AllowCreatePolicyScoped",
"Effect": "Allow",
"Action": [
"iam:CreatePolicy",
"iam:DeletePolicy",
"iam:CreatePolicyVersion",
"iam:DeletePolicyVersion"
],
"Resource": "arn:aws:iam::*:policy/app-*"
},
{
"Sid": "AllowViewIAM",
"Effect": "Allow",
"Action": [
"iam:Get*",
"iam:List*"
],
"Resource": "*"
}
]
}
Step 3: Attach the Boundary
# Create the boundary policy
aws iam create-policy \
--policy-name DeveloperBoundary \
--policy-document file://developer-boundary.json
# Attach boundary to an existing role
aws iam put-role-permissions-boundary \
--role-name developer-role \
--permissions-boundary arn:aws:iam::123456789012:policy/DeveloperBoundary
# Create a new role with boundary
aws iam create-role \
--role-name app-lambda-executor \
--assume-role-policy-document file://trust-policy.json \
--permissions-boundary arn:aws:iam::123456789012:policy/DeveloperBoundary
Step 4: Prevent Privilege Escalation
The boundary must include deny statements to prevent developers from:
- Removing the boundary from their own roles
- Modifying the boundary policy itself
- Creating roles without the boundary attached
- Accessing IAM services to escalate privileges
Step 5: Deploy with Terraform
resource "aws_iam_policy" "developer_boundary" {
name = "DeveloperBoundary"
path = "/"
policy = file("${path.module}/policies/developer-boundary.json")
}
resource "aws_iam_role" "app_role" {
name = "app-lambda-executor"
assume_role_policy = data.aws_iam_policy_document.lambda_trust.json
permissions_boundary = aws_iam_policy.developer_boundary.arn
}
Validation Checklist
- Permission boundary policy created and reviewed by security team
- Boundary includes deny statements preventing self-modification
- Developer delegation policy requires boundary on all new roles
- Role naming convention enforced (e.g.,
app-*prefix) - Developers tested creating roles with and without boundary (should fail without)
- Privilege escalation paths tested and blocked
- CloudTrail logging enabled for IAM API calls
- Boundary policy versioned in source control
- Automated tests validate boundary effectiveness
- Documentation provided to development teams
References
- AWS IAM Permission Boundaries Documentation
- When and Where to Use Permission Boundaries - AWS Security Blog
- AWS Example Permission Boundary - GitHub
- AWS Prescriptive Guidance - Creating Permission Boundaries
Other files in this skill
- LICENSE
- assets/template.md
- references/api-reference.md
- references/standards.md
- references/workflows.md
- scripts/agent.py
- scripts/process.py
assets/template.md (verbatim)
AWS IAM Permission Boundary Implementation Template
Boundary Policy Details
| Field | Value |
|---|---|
| Policy Name | |
| AWS Account ID | |
| Target Roles | (prefix pattern, e.g., app-*) |
| Allowed Services | |
| Created By | |
| Date |
Allowed Service Actions
| Service | Actions Allowed | Justification |
|---|---|---|
| S3 | s3:* | Application data storage |
| DynamoDB | dynamodb:* | Application database |
| Lambda | lambda:* | Serverless compute |
| CloudWatch | cloudwatch:, logs: | Monitoring and logging |
Denied Actions (Escalation Prevention)
| Action | Reason |
|---|---|
| iam:DeleteRolePermissionsBoundary | Prevent boundary removal |
| iam:DeleteUserPermissionsBoundary | Prevent boundary removal |
| iam:CreatePolicyVersion (on boundary) | Prevent boundary modification |
| iam:SetDefaultPolicyVersion (on boundary) | Prevent boundary modification |
Testing Results
| Test Case | Expected Result | Actual Result | Pass/Fail |
|---|---|---|---|
| Create role with boundary | Success | ||
| Create role without boundary | AccessDenied | ||
| Use allowed service (e.g., S3) | Success | ||
| Use blocked service (e.g., IAM admin) | AccessDenied | ||
| Remove own boundary | AccessDenied | ||
| Modify boundary policy | AccessDenied |
Sign-Off
| Role | Name | Date |
|---|---|---|
| Security Engineer | ||
| Cloud Architect | ||
| DevOps Lead |
references/api-reference.md (verbatim)
API Reference: AWS IAM Permission Boundary Agent
Dependencies
| Library | Version | Purpose |
|---|---|---|
| boto3 | >=1.28 | AWS SDK for IAM permission boundary management |
CLI Usage
python scripts/agent.py \
--profile security-admin \
--region us-east-1 \
--audit \
--output-dir /reports/ \
--output iam_boundary_report.json
Functions
get_iam_client(profile, region)
Creates boto3 IAM client with optional named profile.
create_permission_boundary(client, policy_name, allowed_services, allowed_regions) -> dict
Creates an IAM policy for use as a permission boundary. Includes a DenyBoundaryChanges statement to prevent boundary removal. Uses client.create_policy().
attach_boundary_to_role(client, role_name, boundary_arn) -> dict
Calls client.put_role_permissions_boundary() to attach a boundary to a role.
audit_roles_without_boundary(client) -> list
Paginates client.list_roles() and identifies roles missing PermissionsBoundary.
audit_boundary_effectiveness(client, role_name) -> dict
Calls client.get_role(), list_attached_role_policies(), list_role_policies() to show effective policy stack.
generate_report(client) -> dict
Orchestrates audit and generates compliance report.
boto3 IAM Methods Used
| Method | Purpose |
|---|---|
create_policy(PolicyName, PolicyDocument) |
Create boundary policy |
put_role_permissions_boundary(RoleName, PermissionsBoundary) |
Attach boundary |
list_roles() |
Enumerate all roles |
get_role(RoleName) |
Get role details including boundary |
Output Schema
{
"roles_without_boundary_count": 12,
"roles_without_boundary": [{"role_name": "dev-role", "arn": "arn:aws:iam::..."}],
"recommendations": ["Attach permission boundaries to 12 roles"]
}
references/standards.md (verbatim)
AWS IAM Permission Boundaries - Standards Reference
AWS IAM Policy Types
| Policy Type | Scope | Purpose |
|---|---|---|
| Identity-Based | Attached to users/roles/groups | Grants permissions |
| Resource-Based | Attached to resources (S3, KMS) | Cross-account access |
| Permission Boundary | Attached to users/roles | Maximum permission limit |
| Organizations SCP | Attached to OUs/accounts | Organization-wide limit |
| Session Policy | Passed during AssumeRole | Session-level limit |
AWS Well-Architected Framework - Security Pillar
SEC02 - Identity Management
- SEC02-BP02: Use temporary credentials (permission boundaries enforce this)
- SEC02-BP05: Audit and rotate credentials regularly
- SEC02-BP06: Employ user groups and attributes for fine-grained access
SEC03 - Permissions Management
- SEC03-BP01: Define access requirements (boundary defines maximum)
- SEC03-BP02: Grant least privilege access
- SEC03-BP06: Manage access based on lifecycle
- SEC03-BP07: Analyze public and cross-account access
CIS AWS Foundations Benchmark v3.0
- 1.4: Ensure no root account access key exists
- 1.15: Ensure IAM users receive permissions only through groups
- 1.16: Ensure IAM policies that allow full admin privileges are not attached
- 1.17: Ensure a support role has been created for incident management
- 1.22: Ensure IAM policies with admin access are reviewed regularly
NIST SP 800-53 Mapping
- AC-2: Account Management (boundary controls role creation)
- AC-3: Access Enforcement (intersection of policies)
- AC-5: Separation of Duties (boundary prevents security role access)
- AC-6: Least Privilege (boundary enforces maximum permissions)
- AC-6(1): Authorize Access to Security Functions
- AC-6(5): Privileged Accounts (boundary limits even admin roles)
references/workflows.md (verbatim)
AWS IAM Permission Boundaries - Workflows
Boundary Policy Creation Workflow
1. Security team identifies allowed services for developer workloads
│
2. Draft permission boundary policy (JSON)
│
3. Peer review by second security engineer
│
4. Test in sandbox account:
├── Create test role with boundary
├── Verify allowed actions succeed
├── Verify blocked actions are denied
└── Verify boundary cannot be self-modified
│
5. Commit policy to version control (IaC repository)
│
6. Deploy via CI/CD pipeline (Terraform/CloudFormation)
│
7. Attach boundary to all developer-created roles
Developer Role Creation Workflow (with Boundary)
Developer wants to create a new IAM role
│
├── Developer writes role policy (only app-* prefixed)
│
├── Developer creates role with --permissions-boundary flag
│ │
│ └── If boundary not attached → API returns AccessDenied
│
├── AWS IAM validates:
│ ├── Role name matches required prefix (app-*)
│ ├── Permission boundary ARN matches required boundary
│ └── Developer has iam:CreateRole with boundary condition
│
├── Role created successfully with boundary attached
│
└── Effective permissions = identity policy ∩ boundary policy
Privilege Escalation Prevention Workflow
Attacker attempts to escalate privileges:
Attempt 1: Create role without boundary
→ Denied by developer policy (condition requires boundary)
Attempt 2: Modify the boundary policy itself
→ Denied by boundary's own deny statements
Attempt 3: Remove boundary from existing role
→ Denied by boundary deny on DeleteRolePermissionsBoundary
Attempt 4: Create policy granting iam:* access
→ Policy can only grant actions within boundary intersection
Attempt 5: Assume a role without boundary
→ Developer can only create roles with boundary condition
All escalation paths blocked ✓
Boundary Audit Workflow
Monthly audit:
│
├── List all IAM roles in account
│
├── Check each role for boundary attachment:
│ ├── Has boundary → Verify correct boundary ARN
│ └── No boundary → Flag for remediation
│
├── Review boundary policy changes (CloudTrail)
│
├── Check for new IAM actions added to AWS services
│ └── Update boundary if new actions should be restricted
│
└── Generate compliance report
Back to mukul975/Anthropic-Cybersecurity-Skills (817 security skills) or Agent skills.