Security
prowler-sdk-check - Claude MCP Skill
Creates Prowler security checks following SDK architecture patterns. Trigger: When creating or updating a Prowler SDK security check (implementation + metadata) for any provider (AWS, Azure, GCP, K8s, GitHub, etc.).
SEO Guide: Enhance your AI agent with the prowler-sdk-check tool. This Model Context Protocol (MCP) server allows Claude Desktop and other LLMs to creates prowler security checks following sdk architecture patterns. trigger: when creating or updat... Download and configure this skill to unlock new capabilities for your AI workflow.
Documentation
SKILL.md## Check Structure
```text
prowler/providers/{provider}/services/{service}/{check_name}/
āāā __init__.py
āāā {check_name}.py
āāā {check_name}.metadata.json
```
---
## Step-by-Step Creation Process
### 1. Prerequisites
- **Verify check doesn't exist**: Search `prowler/providers/{provider}/services/{service}/`
- **Ensure provider and service exist** - create them first if not
- **Confirm service has required methods** - may need to add/modify service methods to get data
### 2. Create Check Files
```bash
mkdir -p prowler/providers/{provider}/services/{service}/{check_name}
touch prowler/providers/{provider}/services/{service}/{check_name}/__init__.py
touch prowler/providers/{provider}/services/{service}/{check_name}/{check_name}.py
touch prowler/providers/{provider}/services/{service}/{check_name}/{check_name}.metadata.json
```
### 3. Implement Check Logic
```python
from prowler.lib.check.models import Check, Check_Report_{Provider}
from prowler.providers.{provider}.services.{service}.{service}_client import {service}_client
class {check_name}(Check):
"""Ensure that {resource} meets {security_requirement}."""
def execute(self) -> list[Check_Report_{Provider}]:
"""Execute the check logic.
Returns:
A list of reports containing the result of the check.
"""
findings = []
for resource in {service}_client.{resources}:
report = Check_Report_{Provider}(metadata=self.metadata(), resource=resource)
report.status = "PASS" if resource.is_compliant else "FAIL"
report.status_extended = f"Resource {resource.name} compliance status."
findings.append(report)
return findings
```
### 4. Create Metadata File
See complete schema below and `assets/` folder for complete templates.
For detailed field documentation, see `references/metadata-docs.md`.
### 5. Verify Check Detection
```bash
uv run python prowler-cli.py {provider} --list-checks | grep {check_name}
```
### 6. Run Check Locally
```bash
uv run python prowler-cli.py {provider} --log-level ERROR --verbose --check {check_name}
```
### 7. Create Tests
See `prowler-test-sdk` skill for test patterns (PASS, FAIL, no resources, error handling).
---
## Check Naming Convention
```text
{service}_{resource}_{security_control}
```
Examples:
- `ec2_instance_public_ip_disabled`
- `s3_bucket_encryption_enabled`
- `iam_user_mfa_enabled`
---
## Metadata Schema (COMPLETE)
```json
{
"Provider": "aws",
"CheckID": "{check_name}",
"CheckTitle": "Human-readable title",
"CheckType": [
"Software and Configuration Checks/AWS Security Best Practices",
"Software and Configuration Checks/Industry and Regulatory Standards/AWS Foundational Security Best Practices"
],
"ServiceName": "{service}",
"SubServiceName": "",
"ResourceIdTemplate": "",
"Severity": "low|medium|high|critical",
"ResourceType": "AwsEc2Instance|Other",
"ResourceGroup": "security|compute|storage|network",
"Description": "**Bold resource name**. Detailed explanation of what this check evaluates and why it matters.",
"Risk": "What happens if non-compliant. Explain attack vectors, data exposure risks, compliance impact.",
"RelatedUrl": "",
"AdditionalURLs": [
"https://docs.aws.amazon.com/..."
],
"Remediation": {
"Code": {
"CLI": "aws {service} {command} --option value",
"NativeIaC": "```yaml\nResources:\n Resource:\n Type: AWS::{Service}::{Resource}\n Properties:\n Key: value # This line fixes the issue\n```",
"Other": "1. Console steps\n2. Step by step",
"Terraform": "```hcl\nresource \"aws_{service}_{resource}\" \"example\" {\n key = \"value\" # This line fixes the issue\n}\n```"
},
"Recommendation": {
"Text": "Detailed recommendation for remediation.",
"Url": "https://hub.prowler.com/check/{check_name}"
}
},
"Categories": [
"identity-access",
"encryption",
"logging",
"forensics-ready",
"internet-exposed",
"trust-boundaries"
],
"DependsOn": [],
"RelatedTo": [],
"Notes": ""
}
```
### Required Fields
| Field | Description |
|-------|-------------|
| `Provider` | Provider name: aws, azure, gcp, kubernetes, github, m365 |
| `CheckID` | Must match class name and folder name |
| `CheckTitle` | Human-readable title |
| `Severity` | `low`, `medium`, `high`, `critical` |
| `ServiceName` | Service being checked |
| `Description` | What the check evaluates |
| `Risk` | Security impact of non-compliance |
| `Remediation.Code.CLI` | CLI fix command |
| `Remediation.Recommendation.Text` | How to fix |
### Severity Guidelines
| Severity | When to Use |
|----------|-------------|
| `critical` | Direct data exposure, RCE, privilege escalation |
| `high` | Significant security risk, compliance violation |
| `medium` | Defense-in-depth, best practice |
| `low` | Informational, minor hardening |
---
## Check Report Statuses
| Status | When to Use |
|--------|-------------|
| `PASS` | Resource is compliant |
| `FAIL` | Resource is non-compliant |
| `MANUAL` | Requires human verification, or the data needed to evaluate the resource could not be retrieved |
### Permission / availability errors are NOT findings
Never set `FAIL` because an API call failed (missing permission or scope, API not enabled, feature not licensed, data unavailable). That is a scan-configuration problem, not a security issue, and it surfaces as a misleading high-severity finding.
- Service: log the error and expose it distinctly from an empty result (`None` instead of `[]`, an `*_error` attribute, or a `*_lookup_failed` set). Only treat real access errors this way; a `404`/not-found usually means "not configured" and IS a legitimate `FAIL`, and a definitively disabled API is a legitimate `FAIL` when the API's activation is itself the audited control (e.g. GCP Access Approval).
- Check: emit ONE tenant/account/project/subscription-level `MANUAL` finding (not one per resource) whose `status_extended` says the check cannot be evaluated and names the required permission/API/license.
- Do not touch `report.check_metadata.Severity` to hide it.
```python
if <service>_client.<data> is None:
report = CheckReport<Provider>(metadata=self.metadata(), resource={})
report.resource_name = "<Tenant-level resource>"
report.resource_id = "<stable-id>"
report.status = "MANUAL"
report.status_extended = "Cannot evaluate <requirement>: <data> could not be retrieved. Verify that <permission> is granted to the scanning identity."
return [report]
```
---
## Common Patterns
### AWS Check with Regional Resources
```python
from prowler.lib.check.models import Check, Check_Report_AWS
from prowler.providers.aws.services.s3.s3_client import s3_client
class s3_bucket_encryption_enabled(Check):
def execute(self) -> list[Check_Report_AWS]:
findings = []
for bucket in s3_client.buckets.values():
report = Check_Report_AWS(metadata=self.metadata(), resource=bucket)
if bucket.encryption:
report.status = "PASS"
report.status_extended = f"S3 bucket {bucket.name} has encryption enabled."
else:
report.status = "FAIL"
report.status_extended = f"S3 bucket {bucket.name} does not have encryption enabled."
findings.append(report)
return findings
```
### Check with Multiple Conditions
```python
from prowler.lib.check.models import Check, Check_Report_AWS
from prowler.providers.aws.services.ec2.ec2_client import ec2_client
class ec2_instance_hardened(Check):
def execute(self) -> list[Check_Report_AWS]:
findings = []
for instance in ec2_client.instances:
report = Check_Report_AWS(metadata=self.metadata(), resource=instance)
issues = []
if instance.public_ip:
issues.append("has public IP")
if not instance.metadata_options.http_tokens == "required":
issues.append("IMDSv2 not enforced")
if issues:
report.status = "FAIL"
report.status_extended = f"Instance {instance.id} {', '.join(issues)}."
else:
report.status = "PASS"
report.status_extended = f"Instance {instance.id} is properly hardened."
findings.append(report)
return findings
```
---
## Commands
```bash
# Verify detection
uv run python prowler-cli.py {provider} --list-checks | grep {check_name}
# Run check
uv run python prowler-cli.py {provider} --log-level ERROR --verbose --check {check_name}
# Run with specific profile/credentials
uv run python prowler-cli.py aws --profile myprofile --check {check_name}
# Run multiple checks
uv run python prowler-cli.py {provider} --check {check1} {check2} {check3}
```
## Resources
- **Templates**: See [assets/](assets/) for complete check and metadata templates (AWS, Azure, GCP)
- **Documentation**: See [references/metadata-docs.md](references/metadata-docs.md) for official Prowler Developer Guide linksSignals
Information
- Repository
- prowler-cloud/prowler
- Author
- prowler-cloud
- Last Sync
- 9/5/2026
- Repo Updated
- 9/4/2026
- Created
- 1/12/2026
Reviews (0)
No reviews yet. Be the first to review this skill!
Related Skills
upgrade-nodejs
Upgrading Bun's Self-Reported Node.js Version
cursorrules
CrewAI Development Rules
README
Agents ā Working Implementations
cn-check
Install and run the Continue CLI (`cn`) to execute AI agent checks on local code changes. Use when asked to "run checks", "lint with AI", "review my changes with cn", or set up Continue CI locally.
Related Guides
Bear Notes Claude Skill: Your AI-Powered Note-Taking Assistant
Learn how to use the bear-notes Claude skill. Complete guide with installation instructions and examples.
OpenAI Whisper API Claude Skill: Complete Guide to AI-Powered Audio Transcription
Learn how to use the openai-whisper-api Claude skill. Complete guide with installation instructions and examples.
Mastering the Oracle CLI: A Complete Guide to the Claude Skill for Database Professionals
Learn how to use the oracle Claude skill. Complete guide with installation instructions and examples.