@modular-intelligence/cloud-security

v1.0.2

Published

8 days ago

MCP server for multi-cloud security auditing (AWS, Azure, GCP)

0High
0Medium
0Low

Cloud Security MCP Server

A comprehensive multi-cloud security auditing service that integrates with AWS, Azure, and GCP CLIs. This MCP (Model Context Protocol) server enables Claude to audit cloud infrastructure security posture, analyze access controls, review audit logs, and identify security misconfigurations across all major cloud providers.

Overview

This server provides access to security auditing capabilities across three major cloud platforms through a unified read-only interface:

AWS - CloudTrail event searches, GuardDuty threat detection, IAM analysis, S3 bucket exposure checks, security group audits
Azure - Activity log monitoring, resource auditing
GCP - Audit log analysis, project-wide security events

Perfect for cloud security assessments, compliance auditing, incident investigation, access control reviews, and infrastructure hardening.

Tools

| Tool | Cloud Provider | Description | |------|----------------|-------------| | cloudtrail_search | AWS | Search AWS CloudTrail events for security-relevant activity | | guardduty_findings | AWS | List and analyze AWS GuardDuty security findings | | iam_analyze | AWS | Analyze AWS IAM for over-permissioned users and roles | | s3_public_check | AWS | Check S3 bucket for public access exposure | | security_group_audit | AWS | Audit AWS security group rules for overly permissive access | | azure_activity_log | Azure | Search Azure Monitor activity log for events | | gcp_audit_log | GCP | Search GCP audit logs for security events |

CloudTrail Search

Search AWS CloudTrail for security-relevant events including user activity, API calls, and authentication events.

Input Parameters:

{
  event_name: string     // Optional: CloudTrail event name to filter by
  username: string       // Optional: Username to filter by
  region: string         // AWS region (default: us-east-1)
  time_range: string     // Time range: 24h, 7d, or 30d (default: 24h)
  max_results: number    // Maximum results to return 1-100 (default: 20)
}

Example Request:

{
  "event_name": "ConsoleLogin",
  "region": "us-east-1",
  "time_range": "7d",
  "max_results": 50
}

Example Output:

{
  "Events": [
    {
      "EventId": "12345678-1234-1234-1234-123456789012",
      "EventName": "ConsoleLogin",
      "EventTime": "2024-01-15T10:30:45Z",
      "Username": "[email protected]",
      "Resources": [],
      "EventSource": "signin.amazonaws.com",
      "AWSRegion": "us-east-1",
      "CloudTrailEvent": "{\"eventVersion\":\"1.08\",\"eventTime\":\"2024-01-15T10:30:45Z\",\"sourceIPAddress\":\"192.0.2.1\"}"
    },
    {
      "EventId": "87654321-4321-4321-4321-210987654321",
      "EventName": "ConsoleLogin",
      "EventTime": "2024-01-14T15:22:10Z",
      "Username": "[email protected]",
      "Resources": [],
      "EventSource": "signin.amazonaws.com",
      "AWSRegion": "us-east-1",
      "CloudTrailEvent": "{\"eventVersion\":\"1.08\",\"eventTime\":\"2024-01-14T15:22:10Z\",\"sourceIPAddress\":\"203.0.113.5\"}"
    }
  ],
  "NextToken": "eyJFeGVjdXRpb25JZCI6IjEyMzQ1Njc4In0="
}

GuardDuty Findings

List and analyze AWS GuardDuty security findings from threat detection service.

Input Parameters:

{
  detector_id: string    // GuardDuty detector ID (required)
  severity: string       // Optional: Filter by severity (low, medium, high)
  region: string         // AWS region (default: us-east-1)
  max_results: number    // Maximum results to return 1-100 (default: 20)
}

Example Request:

{
  "detector_id": "12b0aac1d12a1e7e0a2b5c8d9e1f2a3b",
  "severity": "high",
  "region": "us-east-1",
  "max_results": 25
}

Example Output:

{
  "Findings": [
    {
      "AccountId": "123456789012",
      "Arn": "arn:aws:guardduty:us-east-1:123456789012:detector/12b0aac1d12a1e7e0a2b5c8d9e1f2a3b/finding/a1b2c3d4e5f6g7h8i9j0k1l2m3n4o5p6",
      "Confidence": 98.0,
      "CreatedAt": "2024-01-15T10:30:45.000Z",
      "Description": "EC2 instance has the IMDSv1 metadata service enabled. This can allow an attacker with code execution to retrieve IAM credentials.",
      "FindingType": "UnauthorizedAccess:EC2/SSHBruteForce",
      "Id": "a1b2c3d4e5f6g7h8i9j0k1l2m3n4o5p6",
      "Partition": "aws",
      "Region": "us-east-1",
      "Resource": {
        "InstanceDetails": {
          "InstanceId": "i-1234567890abcdef0",
          "LaunchTime": "2024-01-10T14:22:33.000Z",
          "NetworkInterfaces": [
            {
              "Ipv6Addresses": [],
              "NetworkInterfaceId": "eni-12345678",
              "PrivateIpAddresses": [
                {
                  "PrivateIpAddress": "172.31.45.123"
                }
              ],
              "PublicDnsName": "ec2-203-0-113-5.compute-1.amazonaws.com",
              "PublicIp": "203.0.113.5",
              "SubnetId": "subnet-12345678",
              "VpcId": "vpc-12345678"
            }
          ]
        }
      },
      "SchemaVersion": "2.0",
      "Severity": 7.5,
      "Title": "EC2 instance i-1234567890abcdef0 has SSH brute force activity detected",
      "UpdatedAt": "2024-01-15T10:30:45.000Z"
    }
  ]
}

IAM Analyze

Analyze AWS IAM users and roles to identify over-permissioned principals and security issues.

Input Parameters:

{
  user_name: string      // Optional: Specific IAM user to analyze
  role_name: string      // Optional: Specific IAM role to analyze
  region: string         // AWS region (default: us-east-1)
}

Example Request:

{
  "user_name": "developer",
  "region": "us-east-1"
}

Example Output:

{
  "users": [
    {
      "UserName": "developer",
      "Arn": "arn:aws:iam::123456789012:user/developer",
      "AttachedPolicies": [
        {
          "PolicyName": "AdministratorAccess",
          "PolicyArn": "arn:aws:iam::aws:policy/AdministratorAccess"
        },
        {
          "PolicyName": "ReadOnlyAccess",
          "PolicyArn": "arn:aws:iam::aws:policy/ReadOnlyAccess"
        }
      ],
      "InlinePolicies": [
        "s3-access-policy",
        "lambda-invoke-policy"
      ],
      "IsOverPermissioned": true,
      "Reasons": [
        "Has AdministratorAccess policy attached",
        "Has 2 attached policies (potential over-permissioning)",
        "Has 2 inline policies (best practice: use managed policies)"
      ]
    }
  ],
  "summary": {
    "totalUsers": 1,
    "overPermissionedUsers": 1,
    "commonIssues": [
      "1 users with excessive permissions detected",
      "1 users have inline policies (use managed policies instead)"
    ]
  }
}

S3 Public Check

Check AWS S3 buckets for public access exposure through ACLs, policies, and access block configuration.

Input Parameters:

{
  bucket_name: string    // S3 bucket name (required, 3-63 chars, lowercase)
  region: string         // AWS region (default: us-east-1)
}

Example Request:

{
  "bucket_name": "my-app-data-bucket",
  "region": "us-east-1"
}

Example Output:

{
  "bucketName": "my-app-data-bucket",
  "isPublic": true,
  "findings": [
    "Bucket policy has 1 statement(s) with wildcard principal",
    "BlockPublicPolicy is not enabled",
    "RestrictPublicBuckets is not enabled"
  ],
  "acl": {
    "Owner": {
      "DisplayName": "owner-name",
      "ID": "79a3abb3738d6cb5be587f8efb2b7387c5cfe7e4c5e5b4c5a8b9c0d1e2f3a4b"
    },
    "Grants": [
      {
        "Grantee": {
          "Type": "Group",
          "URI": "http://acs.amazonaws.com/groups/global/AllUsers"
        },
        "Permission": "READ"
      }
    ]
  },
  "policy": {
    "Version": "2012-10-17",
    "Statement": [
      {
        "Sid": "PublicRead",
        "Effect": "Allow",
        "Principal": "*",
        "Action": "s3:GetObject",
        "Resource": "arn:aws:s3:::my-app-data-bucket/*"
      }
    ]
  },
  "publicAccessBlock": {
    "BlockPublicAcls": true,
    "IgnorePublicAcls": true,
    "BlockPublicPolicy": false,
    "RestrictPublicBuckets": false
  }
}

Security Group Audit

Audit AWS security group rules for overly permissive access patterns and dangerous port exposure.

Input Parameters:

{
  group_id: string       // Optional: Specific security group ID to audit
  region: string         // AWS region (default: us-east-1)
  max_results: number    // Maximum results to return 1-100 (default: 20)
}

Example Request:

{
  "region": "us-east-1",
  "max_results": 10
}

Example Output:

{
  "summary": {
    "totalGroups": 2,
    "compliantGroups": 1,
    "nonCompliantGroups": 1,
    "totalFindings": 2
  },
  "audits": [
    {
      "groupId": "sg-12345678",
      "groupName": "default",
      "vpcId": "vpc-12345678",
      "findings": [
        "CRITICAL: Rule allows public access (0.0.0.0/0) to sensitive port(s) 22-22/tcp",
        "HIGH: Rule allows public access (0.0.0.0/0) to wide port range 443-8443/tcp"
      ],
      "riskyRules": [
        {
          "IpProtocol": "tcp",
          "FromPort": 22,
          "ToPort": 22,
          "IpRanges": [
            {
              "CidrIp": "0.0.0.0/0",
              "Description": "SSH access from anywhere"
            }
          ]
        }
      ],
      "isCompliant": false
    },
    {
      "groupId": "sg-87654321",
      "groupName": "app-server-sg",
      "vpcId": "vpc-87654321",
      "findings": [
        "No risky rules detected"
      ],
      "riskyRules": [],
      "isCompliant": true
    }
  ]
}

Azure Activity Log

Search Azure Monitor activity log for events and activities across Azure resources.

Input Parameters:

{
  resource_group: string // Optional: Resource group to filter by
  time_range: string     // Time range: 24h, 7d, or 30d (default: 24h)
  subscription: string   // Optional: Azure subscription ID
  max_results: number    // Maximum results to return 1-100 (default: 20)
}

Example Request:

{
  "resource_group": "prod-resources",
  "time_range": "7d",
  "max_results": 30
}

Example Output:

[
  {
    "eventTimestamp": "2024-01-15T10:30:45.123Z",
    "id": "/subscriptions/12345678-1234-1234-1234-123456789012/resourceGroups/prod-resources/providers/Microsoft.Authorization/policyAssignments/12345678-1234-1234-1234-123456789012/events/12345678-1234-1234-1234-123456789012/ticks/638420346450000000",
    "resourceId": "/subscriptions/12345678-1234-1234-1234-123456789012/resourceGroups/prod-resources/providers/Microsoft.Network/networkSecurityGroups/prod-nsg",
    "resourceGroupName": "prod-resources",
    "resourceProvider": {
      "value": "Microsoft.Network",
      "localizedValue": "Microsoft.Network"
    },
    "resourceType": {
      "value": "Microsoft.Network/networkSecurityGroups",
      "localizedValue": "Microsoft.Network/networkSecurityGroups"
    },
    "operationName": {
      "value": "Microsoft.Network/networkSecurityGroups/write",
      "localizedValue": "Write Network Security Group"
    },
    "status": {
      "value": "Succeeded",
      "localizedValue": "Succeeded"
    },
    "subStatus": {
      "value": "Created",
      "localizedValue": "Created"
    },
    "caller": "[email protected]",
    "correlationId": "12345678-1234-1234-1234-123456789012",
    "eventSource": {
      "value": "Administrative",
      "localizedValue": "Administrative"
    },
    "level": {
      "value": "Informational",
      "localizedValue": "Informational"
    }
  }
]

GCP Audit Log

Search GCP audit logs for security and compliance events across projects and services.

Input Parameters:

{
  project: string        // Optional: GCP project ID
  log_filter: string     // Optional: Additional log filter expression
  time_range: string     // Time range: 24h, 7d, or 30d (default: 24h)
  max_results: number    // Maximum results to return 1-100 (default: 20)
}

Example Request:

{
  "project": "my-gcp-project",
  "log_filter": "protoPayload.methodName=compute.instances.delete",
  "time_range": "30d",
  "max_results": 15
}

Example Output:

[
  {
    "insertId": "12345678-1234-1234-1234-123456789012",
    "logName": "projects/my-gcp-project/logs/cloudaudit.googleapis.com%2Factivity",
    "protoPayload": {
      "methodName": "compute.instances.delete",
      "resourceName": "projects/my-gcp-project/zones/us-central1-a/instances/web-server-1",
      "request": {
        "@type": "type.googleapis.com/compute.instances.delete"
      },
      "response": {
        "@type": "type.googleapis.com/operation"
      },
      "status": {
        "code": 0
      },
      "authenticationInfo": {
        "principalEmail": "[email protected]"
      },
      "requestMetadata": {
        "callerIp": "192.0.2.1",
        "userAgent": "gcloud-cli"
      },
      "serviceName": "compute.googleapis.com"
    },
    "receiveTimestamp": "2024-01-15T10:30:45.123Z",
    "severity": "NOTICE",
    "timestamp": "2024-01-15T10:30:44.123Z"
  },
  {
    "insertId": "87654321-4321-4321-4321-210987654321",
    "logName": "projects/my-gcp-project/logs/cloudaudit.googleapis.com%2Factivity",
    "protoPayload": {
      "methodName": "storage.buckets.setIamPolicy",
      "resourceName": "projects/_/buckets/my-data-bucket",
      "request": {
        "@type": "type.googleapis.com/storage.buckets.setIamPolicy"
      },
      "authenticationInfo": {
        "principalEmail": "[email protected]"
      },
      "requestMetadata": {
        "callerIp": "203.0.113.5"
      },
      "serviceName": "storage.googleapis.com"
    },
    "receiveTimestamp": "2024-01-14T15:22:10.456Z",
    "severity": "NOTICE",
    "timestamp": "2024-01-14T15:22:09.456Z"
  }
]

Configuration

AWS CLI Authentication

This server requires the AWS CLI to be configured with appropriate credentials.

Setup AWS CLI:

aws configure

Or use environment variables:

export AWS_ACCESS_KEY_ID="your-access-key"
export AWS_SECRET_ACCESS_KEY="your-secret-key"
export AWS_DEFAULT_REGION="us-east-1"

For detailed instructions, see: https://docs.aws.amazon.com/cli/latest/userguide/cli-configure-quickstart.html

Azure CLI Authentication

Configure the Azure CLI with your credentials.

Login to Azure:

az login

For more information, see: https://learn.microsoft.com/en-us/cli/azure/authenticate-azure-cli

GCP CLI Authentication

Set up Google Cloud CLI authentication.

Login to GCP:

gcloud auth login
gcloud config set project YOUR_PROJECT_ID

For detailed instructions, see: https://cloud.google.com/docs/authentication/gcloud

Prerequisites

Bun runtime (version 1.x or later) or Node.js 18+
AWS CLI (for AWS tools)
Azure CLI (for Azure tools)
Google Cloud CLI (for GCP tools)
Appropriate cloud credentials configured for each CLI

Installation Commands

AWS CLI:

curl "https://awscli.amazonaws.com/awscli-exe-linux-x86_64.zip" -o "awscliv2.zip"
unzip awscliv2.zip
sudo ./aws/install

Azure CLI:

curl -sL https://aka.ms/InstallAzureCLIDeb | sudo bash

Google Cloud CLI:

curl https://sdk.cloud.google.com | bash
exec -l $SHELL
gcloud init

Installation

Prerequisites

Bun runtime (version 1.x or later)
Node.js 18+ (alternative runtime)
AWS CLI configured with credentials (for AWS tools)
Azure CLI configured with credentials (for Azure tools)
Google Cloud CLI configured with credentials (for GCP tools)

Steps

Clone or download this repository:

git clone <repo-url>
cd cloud-security

Install dependencies:

bun install

Build the project:

bun run build

Ensure cloud CLIs are authenticated:

aws configure          # Configure AWS CLI
az login              # Login to Azure
gcloud auth login     # Login to GCP

Run the server:

bun run start

The server will start listening on stdio transport.

Usage

Running the Server

Start the server with Bun:

bun run src/index.ts

The server implements the Model Context Protocol (MCP) and communicates via stdio transport. It can be integrated with Claude or other MCP clients.

Claude Desktop Configuration

Add the server to your Claude Desktop configuration at ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "cloud-security": {
      "command": "bun",
      "args": [
        "run",
        "/path/to/cloud-security/src/index.ts"
      ]
    }
  }
}

Claude Code MCP Settings

Configure the server in Claude Code's MCP settings (typically in .mcp.json or via settings UI):

{
  "servers": {
    "cloud-security": {
      "transport": "stdio",
      "command": "bun",
      "args": ["run", "/path/to/cloud-security/src/index.ts"]
    }
  }
}

Example Usage in Claude

Once configured, you can use the tools directly in conversations with Claude:

Request: "Audit my AWS security groups in us-east-1 for overly permissive rules"

Claude will call:

{
  "tool": "security_group_audit",
  "input": {
    "region": "us-east-1",
    "max_results": 50
  }
}

Request: "Check if my S3 bucket 'company-data' is publicly accessible"

Claude will call:

{
  "tool": "s3_public_check",
  "input": {
    "bucket_name": "company-data",
    "region": "us-east-1"
  }
}

Request: "Search CloudTrail for failed login attempts in the last 7 days"

Claude will call:

{
  "tool": "cloudtrail_search",
  "input": {
    "event_name": "ConsoleLogin",
    "region": "us-east-1",
    "time_range": "7d",
    "max_results": 100
  }
}

Request: "Analyze IAM users in my AWS account to find over-permissioned principals"

Claude will call:

{
  "tool": "iam_analyze",
  "input": {
    "region": "us-east-1"
  }
}

Request: "Get high severity GuardDuty findings from the last 30 days"

Claude will call:

{
  "tool": "guardduty_findings",
  "input": {
    "detector_id": "12b0aac1d12a1e7e0a2b5c8d9e1f2a3b",
    "severity": "high",
    "region": "us-east-1"
  }
}

Security

This server implements strict read-only enforcement and comprehensive input validation to prevent unauthorized cloud modifications and injection attacks.

Read-Only Enforcement

All tools are strictly read-only and implement action-blocking at multiple levels:

AWS CLI Blocking:

Blocked actions: delete, remove, terminate, modify, update, create, put, attach, detach, enable, disable, revoke, deregister, stop, start, reboot
Only lookup, list, describe, and get operations allowed
Examples of prevented operations:
- Cannot delete instances, buckets, or security groups
- Cannot modify IAM policies or users
- Cannot create or update security groups
- Cannot terminate EC2 instances

Azure CLI Blocking:

Blocked actions: delete, create, update, set, start, stop, restart
Only list and show operations allowed
Prevents modification of Azure resources

GCP CLI Blocking:

Blocked actions: delete, create, update, set, start, stop, deploy
Only list and describe operations allowed
Prevents creation or modification of GCP resources

Input Validation

Region Validation (AWS)

Validates against whitelist of official AWS regions
Rejects invalid or custom region strings
Supported regions: us-east-1, us-east-2, us-west-1, us-west-2, eu-west-1, eu-west-2, eu-west-3, eu-central-1, ap-southeast-1, ap-southeast-2, ap-northeast-1, ap-northeast-2, ap-south-1, sa-east-1, ca-central-1, me-south-1, af-south-1

S3 Bucket Name Validation

Requires RFC 5321 compliant bucket names
Minimum length: 3 characters
Maximum length: 63 characters
Character set: lowercase letters, numbers, dots, hyphens
Must start and end with alphanumeric character

Resource ID Validation

Alphanumeric characters, colons, slashes, hyphens only
Maximum length: 500 characters
Prevents command injection via resource identifiers

CLI Argument Validation

Blocks shell metacharacters: semicolons, pipes, ampersands, backticks, dollar signs
Prevents shell injection attacks
Validates against malicious operator chaining

What Gets Blocked

The server rejects:

Write/delete/modify operations on AWS, Azure, or GCP resources
Invalid or out-of-range AWS region specifications
Malformed S3 bucket names
Invalid resource identifiers with special characters
Arguments containing shell metacharacters
Oversized input parameters

Error Handling

Invalid inputs return descriptive error messages
CLI errors are caught and reported
Missing CLI tools trigger helpful configuration messages
Command timeouts are handled gracefully (120-second default)
Large output buffers (up to 10MB) prevent truncation

License

ISC License - see LICENSE file for details

Published

Vulnerabilities

Links

Maintainers

Keywords

Readme

Cloud Security MCP Server

Overview

Tools

CloudTrail Search

GuardDuty Findings

IAM Analyze

S3 Public Check

Security Group Audit

Azure Activity Log

GCP Audit Log

Configuration

AWS CLI Authentication

Azure CLI Authentication

GCP CLI Authentication

Prerequisites

Installation Commands

Installation

Prerequisites

Steps

Usage

Running the Server

Claude Desktop Configuration

Claude Code MCP Settings

Example Usage in Claude

Security

Read-Only Enforcement

Input Validation

What Gets Blocked

Error Handling

License