add prowler v4 parser

docs/content/en/integrations/parsers/file/aws_prowler_v4.md

@@ -0,0 +1,102 @@
+---
+title: "AWS Prowler V3"
+toc_hide: true
+---
+
+### File Types
+DefectDojo parser accepts a .json-ocsf file. Please note: earlier versions of AWS Prowler create output data in a different format. See our other documentation if you are using an earlier version of AWS Prowler: 
+* [Prowler v2](https://documentation.defectdojo.com/integrations/parsers/file/aws_prowler/) 
+* [Prowler v3](https://documentation.defectdojo.com/integrations/parsers/file/aws_prowler_v3/) 
+
+JSON reports can be created from the [AWS Prowler V4 CLI](https://docs.prowler.cloud/en/latest/tutorials/reporting/#json) using the following command: `prowler <provider> -M json-ocsf`
+
+### Acceptable JSON Format
+The parser expects an array of assessments. All properties are strings and are required by the parser.
+
+~~~
+
+[{
+    "metadata": {
+        "event_code": "iam_role_administratoraccess_policy_permissive_trust_relationship",
+        "product": {
+            "name": "Prowler",
+            "vendor_name": "Prowler",
+            "version": "4.2.1"
+        },
+        "version": "1.2.0"
+    },
+    "severity_id": 4,
+    "severity": "High",
+    "status": "Suppressed",
+    "status_code": "FAIL",
+    "status_detail": "IAM Role myAdministratorExecutionRole has AdministratorAccess policy attached that has too permissive trust relationship.",
+    "status_id": 3,
+    "unmapped": {
+        "check_type": "",
+        "related_url": "https://docs.aws.amazon.com/IAM/latest/UserGuide/access_policies_job-functions.html#jf_administrator",
+        "categories": "trustboundaries",
+        "depends_on": "",
+        "related_to": "",
+        "notes": "CAF Security Epic: IAM",
+        "compliance": {}
+    },
+    "activity_name": "Create",
+    "activity_id": 1,
+    "finding_info": {
+        "created_time": "2024-06-03T14:15:19.382075",
+        "desc": "Ensure IAM Roles with attached AdministratorAccess policy have a well defined trust relationship",
+        "product_uid": "prowler",
+        "title": "Ensure IAM Roles with attached AdministratorAccess policy have a well defined trust relationship",
+        "uid": "prowler-aws-iam_role_administratoraccess_policy_permissive_trust_relationship-123456789012-us-east-1-myAdministratorExecutionRole"
+    },
+    "resources": [
+        {
+            "cloud_partition": "aws",
+            "region": "us-east-1",
+            "data": {
+                "details": ""
+            },
+            "group": {
+                "name": "iam"
+            },
+            "labels": [],
+            "name": "myAdministratorExecutionRole",
+            "type": "AwsIamRole",
+            "uid": "arn:aws:iam::123456789012:role/myAdministratorExecutionRole"
+        }
+    ],
+    "category_name": "Findings",
+    "category_uid": 2,
+    "class_name": "DetectionFinding",
+    "class_uid": 2004,
+    "cloud": {
+        "account": {
+            "name": "",
+            "type": "AWS_Account",
+            "type_id": 10,
+            "uid": "123456789012",
+            "labels": []
+        },
+        "org": {
+            "name": "",
+            "uid": ""
+        },
+        "provider": "aws",
+        "region": "us-east-1"
+    },
+    "event_time": "2024-06-03T14:15:19.382075",
+    "remediation": {
+        "desc": "Apply the principle of least privilege. Instead of AdministratorAccess, assign only the permissions necessary for specific roles and tasks. Create custom IAM policies with minimal permissions based on the principle of least privilege. If a role really needs AdministratorAccess, the trust relationship must be well defined to restrict it usage only to the Principal, Action, Audience and Subject intended for it.",
+        "references": [
+            "https://docs.aws.amazon.com/IAM/latest/UserGuide/best-practices.html#grant-least-privilege"
+        ]
+    },
+    "risk_details": "The AWS-managed AdministratorAccess policy grants all actions for all AWS services and for all resources in the account and as such exposes the customer to a significant data leakage threat. It is therefore particularly important that the trust relationship is well defined to restrict it usage only to the Principal, Action, Audience and Subject intended for it.",
+    "type_uid": 200401,
+    "type_name": "Create"
+}]
+
+~~~
+
+### Sample Scan Data
+Unit tests of AWS Prowler V4 JSON-OCSF can be found at https://github.com/DefectDojo/django-DefectDojo/tree/master/unittests/scans/aws_prowler_v4.

dojo/tools/aws_prowler_v4/parser.py

@@ -0,0 +1,108 @@
+
+import hashlib
+import json
+import textwrap
+from datetime import date
+
+from dojo.models import Finding
+
+
+class AWSProwlerV4Parser:
+    SCAN_TYPE = ["AWS Prowler V4"]
+
+    def get_scan_types(self):
+        return AWSProwlerV4Parser.SCAN_TYPE
+
+    def get_label_for_scan_types(self, scan_type):
+        return AWSProwlerV4Parser.SCAN_TYPE[0]
+
+    def get_description_for_scan_types(self, scan_type):
+        return "Export of AWS Prowler V4 JSON OCSF v1.1.0 format."
+
+    def get_findings(self, file, test):
+        if file.name.lower().endswith('.json'):
+            return self.process_json(file, test)
+        else:
+            msg = 'Unknown file format'
+            raise ValueError(msg)
+
+    def process_json(self, file, test):
+        dupes = {}
+
+        data = json.load(file)
+        # mapping of json fields between Prowler v3 and v4:
+        # https://docs.prowler.com/projects/prowler-open-source/en/latest/tutorials/reporting/#json
+        for deserialized in data:
+
+            status = deserialized.get('status_code')
+            if status.upper() != 'FAIL':
+                continue
+
+            account_id = deserialized.get('cloud', {}).get('account', {}).get("uid", '')
+            region = deserialized.get('resources', [{}])[0].get('region', '')
+            provider = deserialized.get('cloud', {}).get('provider', '')
+            compliance = ''
+            compliance_field = deserialized.get('unmapped', {}).get("compliance", {})
+            if compliance_field:
+                compliance = ' | '.join([f"{key}:{','.join(value)}" for key, value in compliance_field.items()])
+            result_extended = deserialized.get('status_detail')
+            general_description = deserialized.get('finding_info', {}).get('desc', '')
+            asff_compliance_type = deserialized.get('unmapped', {}).get('check_type', '')
+            severity = deserialized.get('severity', 'Info').capitalize()
+            aws_service_name = deserialized.get('resources', [{}])[0].get('group', {}).get('name', '')
+            impact = deserialized.get('risk_details')
+            mitigation = deserialized.get('remediation', {}).get("desc", '')
+            documentation = deserialized.get('remediation', {}).get("references", '')
+            documentation = str(documentation) + "\n" + str(deserialized.get('unmapped', {}).get('related_url', ''))
+            security_domain = deserialized.get('resources', [{}])[0].get('type', '')
+            timestamp = deserialized.get("event_time")
+            resource_arn = deserialized.get('resources', [{}])[0].get('uid', '')
+            resource_id = deserialized.get('resources', [{}])[0].get('name', '')
+            unique_id_from_tool = deserialized.get('finding_info', {}).get('uid', '')
+            if not resource_arn or resource_arn == "":
+                component_name = str(provider) + "-" + str(account_id) + "-" + str(region) + "-" + str(resource_id)
+            else:
+                component_name = resource_arn
+
+            description = "**Issue:** " + str(result_extended) + \
+                "\n**Description:** " + str(general_description) + \
+                "\n**AWS Account:** " + str(account_id) + \
+                "\n**Region:** " + str(region) + \
+                "\n**AWS Service:** " + str(aws_service_name) + \
+                "\n**Security Domain:** " + str(security_domain) + \
+                "\n**Compliance:** " + str(compliance) + \
+                "\n**ASFF Compliance Type:** " + str(asff_compliance_type)
+
+            # improving key to get duplicates
+            dupe_key = hashlib.sha256(unique_id_from_tool.encode('utf-8')).hexdigest()
+            if dupe_key in dupes:
+                find = dupes[dupe_key]
+                if description is not None:
+                    find.description += description + "\n\n"
+                find.nb_occurences += 1
+            else:
+                find = Finding(
+                    title=textwrap.shorten(result_extended, 150),
+                    cwe=1032,  # Security Configuration Weaknesses, would like to fine tune
+                    test=test,
+                    description=description,
+                    component_name=component_name,
+                    unique_id_from_tool=unique_id_from_tool,
+                    severity=severity,
+                    references=documentation,
+                    date=date.fromisoformat(timestamp[:10]),
+                    static_finding=True,
+                    dynamic_finding=False,
+                    nb_occurences=1,
+                    mitigation=mitigation,
+                    impact=impact,
+                )
+                dupes[dupe_key] = find
+
+        return list(dupes.values())
+
+    def formatview(self, depth):
+        if depth > 1:
+            return "* "
+        else:
+            return ""