Dataset Preview
Duplicate
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed
Error code:   DatasetGenerationError
Exception:    CastError
Message:      Couldn't cast
id: string
side: string
category: string
source: string
type: string
state: string
question: string
candidates: struct<true_positive: string, benign_positive: string, false_positive: string>
  child 0, true_positive: string
  child 1, benign_positive: string
  child 2, false_positive: string
target: string
criteria: struct<true: string, false: string>
  child 0, true: string
  child 1, false: string
to
{'id': Value('string'), 'side': Value('string'), 'category': Value('string'), 'type': Value('string'), 'state': Value('string'), 'question': Value('string'), 'criteria': {'true': Value('string'), 'false': Value('string')}, 'target': Value('bool')}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1835, in _prepare_split_single
                  for key, table in generator:
                                    ^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 621, in wrapped
                  for item in generator(*args, **kwargs):
                              ~~~~~~~~~^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 361, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2381, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2309, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              id: string
              side: string
              category: string
              source: string
              type: string
              state: string
              question: string
              candidates: struct<true_positive: string, benign_positive: string, false_positive: string>
                child 0, true_positive: string
                child 1, benign_positive: string
                child 2, false_positive: string
              target: string
              criteria: struct<true: string, false: string>
                child 0, true: string
                child 1, false: string
              to
              {'id': Value('string'), 'side': Value('string'), 'category': Value('string'), 'type': Value('string'), 'state': Value('string'), 'question': Value('string'), 'criteria': {'true': Value('string'), 'false': Value('string')}, 'target': Value('bool')}
              because column names don't match
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1379, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      builder, max_dataset_size_bytes=max_dataset_size_bytes
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 956, in stream_convert_to_parquet
                  builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
                  ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1702, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ~~~~~~~~~~~~~~~~~~~~~~~~~~^
                      gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  ):
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1887, in _prepare_split_single
                  raise DatasetGenerationError("An error occurred while generating the dataset") from e
              datasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

id
string
side
string
category
string
type
string
state
string
question
string
criteria
dict
target
bool
cld-029
offensive
off_scanner_adjudication
boolean
Terraform aws_iam_role_policy_attachment binds AdministratorAccess to the Lambda execution role "image-resizer". A thumbnail function runs with full admin.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-057
offensive
off_scanner_adjudication
boolean
Terraform aws_s3_bucket_policy: {"Effect":"Allow","Principal":"*","Action":"s3:*","Resource":["arn:aws:s3:::assets","arn:aws:s3:::assets/*"]}. Anonymous users get full bucket control including delete.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-091
offensive
off_scanner_adjudication
boolean
Terraform aws_instance "bastion" iam_instance_profile points to a role with AdministratorAccess. A compromised SSRF on this host yields full account admin via the metadata service.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-241
offensive
off_scanner_adjudication
boolean
IAM role 'app-runtime' policy grants iam:PassRole on Resource '*', allowing it to pass any role including admin roles to AWS services.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-ck-016
offensive
off_scanner_adjudication
boolean
Checkov CKV_AWS_189 FAILED on aws_ebs_volume.secrets_vol: Ensure EBS volume is encrypted by KMS using a customer managed key.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
true
cloud-chai-006
offensive
off_chain_worthy
boolean
GCP Cloud Storage bucket name follows pattern 'project-backups-YYYY'. Bucket is not publicly accessible. Name disclosure alone does not provide access.
Is this finding worth chaining with another for higher impact?
{ "true": "chains to a meaningful escalation (ATO, data access, lateral movement)", "false": "standalone only — no useful chain target or the chain is theoretical" }
false
cld-ck-026
offensive
off_scanner_adjudication
boolean
Checkov CKV_AWS_30 FAILED on aws_elasticache_replication_group.redis_transit: Ensure all data stored in the Elasticache Replication Group is securely encrypted in transit.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
true
cld-ck-040
offensive
off_scanner_adjudication
boolean
Checkov CKV_AWS_108 FAILED on aws_iam_policy.exfil: Ensure IAM policies does not allow data exfiltration.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
true
cld-ck-093
offensive
off_scanner_adjudication
boolean
Checkov CKV2_AWS_16 FAILED on aws_autoscaling_group.web: Ensure that Auto Scaling groups associated with a load balancer use the ELB health checks.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
false
cld-027
offensive
off_scanner_adjudication
boolean
Terraform aws_efs_file_system "shared": no encrypted argument set (defaults to false in this provider version). The shared file system is unencrypted.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-ck-010
offensive
off_scanner_adjudication
boolean
Checkov CKV_AWS_70 FAILED on aws_s3_bucket.shared: Ensure S3 bucket does not allow an action with any Principal.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
true
cld-074
offensive
off_scanner_adjudication
boolean
Terraform aws_s3_bucket_policy enforcing TLS: {"Effect":"Deny","Principal":"*","Action":"s3:*","Resource":["arn:aws:s3:::data","arn:aws:s3:::data/*"],"Condition":{"Bool":{"aws:SecureTransport":"false"}}}. A deny-insecure-transport guardrail.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
false
cld-ck-021
offensive
off_scanner_adjudication
boolean
Checkov CKV_AWS_26 FAILED on aws_sns_topic.alerts: Ensure all data stored in the SNS topic is encrypted.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
true
cld-206
offensive
off_scanner_adjudication
boolean
AWS security group sg-0f6 allows inbound TCP 27017 from 0.0.0.0/0 to a MongoDB replica.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-ck-096
offensive
off_scanner_adjudication
boolean
Checkov CKV_DOCKER_7 FAILED on Dockerfile.base: Ensure the base image uses a non latest version tag.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
false
cld-109
offensive
off_scanner_adjudication
boolean
Terraform aws_lambda_permission { principal = "*", action = "lambda:InvokeFunction" } on a function that processes payments. Any caller can invoke it.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-254
offensive
off_scanner_adjudication
boolean
RDS MySQL instance 'acme-billing' has StorageEncrypted=false; it stores payment records.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-147
offensive
off_scanner_adjudication
boolean
aws_security_group ingress { from_port=5432 to_port=5432 cidr_blocks=["198.51.100.10/32"] } — Postgres restricted to a single management host.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
false
cld-ms-079
offensive
off_scanner_adjudication
boolean
The worker Lambda is not placed inside the VPC, so it cannot reach private subnets.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
false
cloud-verd-002
offensive
off_verdict
boolean
HTTP 403 from http://169.254.169.254/latest/meta-data/ via SSRF. Response: '<?xml version="1.0" encoding="UTF-8"?><Error><Code>AccessDenied</Code></Error>'. IMDSv2 is enforced — token hop limit blocks SSRF.
Is the probe result VULNERABLE?
{ "true": "the evidence confirms exploitability", "false": "the evidence is inconclusive, a false positive, or the control is effective" }
false
cld-135
offensive
off_scanner_adjudication
boolean
Terraform aws_neptune_cluster { storage_encrypted = false } and an SG allowing 8182 from 0.0.0.0/0. The graph DB is unencrypted and internet-reachable.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-252
offensive
off_scanner_adjudication
boolean
An IAM permission boundary is attached to all developer roles, capping effective permissions to a defined allowlist.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
false
cld-ck-087
offensive
off_scanner_adjudication
boolean
Checkov CKV_AWS_338 FAILED on aws_cloudwatch_log_group.app_logs: Ensure CloudWatch log groups retains logs for at least 1 year.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
false
cld-ck-047
offensive
off_scanner_adjudication
boolean
Checkov CKV_AWS_64 FAILED on aws_redshift_cluster.warehouse: Ensure all data stored in the Redshift cluster is securely encrypted at rest.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
true
cld-028
offensive
off_scanner_adjudication
boolean
Terraform aws_lambda_function environment { variables = { DB_HOST_SSM = "/prod/db/host", LOG_LEVEL = "INFO" } }; the password is fetched at runtime from SSM, not stored in env.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
false
cld-117
offensive
off_scanner_adjudication
boolean
boto3 code: iam.put_user_policy(UserName="svc-reporting", PolicyDocument=json.dumps({"Version":"2012-10-17","Statement":[{"Effect":"Allow","Action":"*","Resource":"*"}])). Grants the service user admin.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-216
offensive
off_scanner_adjudication
boolean
AWS security group sg-1c3 on an internet-facing web tier allows inbound TCP 80 and TCP 443 from 0.0.0.0/0.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
false
cld-223
offensive
off_scanner_adjudication
boolean
Azure NSG rule 'allow-https-in' permits inbound TCP 443 from 'Internet' to an Application Gateway subnet.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
false
cld-ms-036
offensive
off_scanner_adjudication
boolean
prowler check elbv2_cross_zone_load_balancing_enabled: FAIL — ALB internal-svc does not have cross-zone balancing enabled.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
false
cld-226
offensive
off_scanner_adjudication
boolean
S3 bucket policy on 'acme-customer-exports' grants s3:GetObject to Principal '*' with no condition; the bucket stores CSV exports of customer PII.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-128
offensive
off_scanner_adjudication
boolean
Terraform aws_elasticache_replication_group { at_rest_encryption_enabled = true, transit_encryption_enabled = true, auth_token = var.redis_auth } with an SG limited to the app SG. Encrypted, authenticated Redis.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
false
cld-018
offensive
off_scanner_adjudication
boolean
Terraform aws_db_instance "analytics" { storage_encrypted = true, kms_key_id = aws_kms_key.rds.arn }. RDS is encrypted at rest with a customer key.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
false
cld-287
offensive
off_scanner_adjudication
boolean
Azure Monitor Activity Log export and Microsoft Defender for Cloud are both disabled subscription-wide; no audit trail of control-plane changes is retained or alerted on.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-243
offensive
off_scanner_adjudication
boolean
IAM role 'vendor-access' trust policy allows a third-party AWS account to AssumeRole but omits the sts:ExternalId condition (confused-deputy exposure).
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cloud-temp-006
offensive
off_template_relevance
boolean
Target: S3 static website hosting. Template: 'sqli-union-based'. Tests SQL injection — no database backend exists on a static S3 site.
Is this probe template relevant to the target?
{ "true": "the template tests a class that applies to this tech stack", "false": "wrong technology, wrong version, or the class cannot exist on this stack" }
false
cloud-aler-001
defensive
def_alert_triage
boolean
GuardDuty finding: UnauthorizedAccess:IAMUser/InstanceCredentialExfiltration. IAM credentials from EC2 instance i-0abc123 used from external IP 45.33.32.156 (not the instance IP). Credential type: temporary role credentials.
Is this security alert a true positive?
{ "true": "real malicious or unauthorized activity", "false": "legitimate activity, misconfigured rule, or known benign behavior" }
true
cld-ck-044
offensive
off_scanner_adjudication
boolean
Checkov CKV_AWS_79 FAILED on aws_instance.web: Ensure Instance Metadata Service Version 1 is not enabled.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
true
cloud-aler-002
defensive
def_alert_triage
boolean
CloudTrail: ConsoleLogin from IP 203.0.113.50, user: admin@company.com, MFA: false, sourceIPAddress: 203.0.113.50 (Tor exit node). Event time: 03:42 UTC (outside business hours).
Is this security alert a true positive?
{ "true": "real malicious or unauthorized activity", "false": "legitimate activity, misconfigured rule, or known benign behavior" }
true
cld-069
offensive
off_scanner_adjudication
boolean
Lambda handler (Python): client = boto3.client("rds", aws_access_key_id="AKIA2E0A8F3B1C4D5E6F", aws_secret_access_key="hV8...redacted..."). Static IAM creds are embedded in function source.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-296
offensive
off_scanner_adjudication
boolean
EC2 AMI 'ami-0acme' is shared publicly (LaunchPermission 'all'); the image contains a baked-in application with embedded credentials.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-314
offensive
off_scanner_adjudication
boolean
AWS ECR repository 'acme/app' policy grants pull only to specific account principals and has scan-on-push enabled.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
false
cld-126
offensive
off_scanner_adjudication
boolean
Terraform aws_mq_broker { publicly_accessible = false } in private subnets with a generated password stored in Secrets Manager. The broker is internal only.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
false
cld-250
offensive
off_scanner_adjudication
boolean
Azure role assignment grants the 'Reader' role scoped to a single resource group to a monitoring service principal.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
false
cld-282
offensive
off_scanner_adjudication
boolean
The application reads its database credentials from AWS Systems Manager Parameter Store (SecureString) at startup; no credentials appear in source.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
false
cld-ms-009
offensive
off_scanner_adjudication
boolean
tfsec aws-ecs-enable-container-insight (LOW): ECS cluster does not have Container Insights enabled. Resource aws_ecs_cluster.main.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
false
cld-ck-085
offensive
off_scanner_adjudication
boolean
Checkov CKV_AWS_23 FAILED on aws_security_group.app_sg: Ensure every security group and rule has a description.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
false
cld-131
offensive
off_scanner_adjudication
boolean
Terraform aws_apigatewayv2_api with default_route_settings and a $default route whose authorization_type is "NONE" on an API that mutates account data.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-099
offensive
off_scanner_adjudication
boolean
Lambda env (serverless.yml): environment: { AWS_SECRET_ACCESS_KEY: "wJalrXUtnFEMI/bPxRfiCYEXAMPLEKEY", GITHUB_TOKEN: "ghp_16C7...redacted" }. Credentials are committed in the deploy config.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-ck-089
offensive
off_scanner_adjudication
boolean
Checkov CKV_AWS_67 FAILED on aws_cloudtrail.main: Ensure CloudTrail log file validation is enabled.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
false
cld-ms-051
offensive
off_scanner_adjudication
boolean
Snyk SNYK-CC-00119 (Low): RDS instance has backup retention set below the recommended value for a non-production database.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
false
cld-ck-053
offensive
off_scanner_adjudication
boolean
Checkov CKV_AWS_115 FAILED on aws_lambda_function.worker: Ensure that AWS Lambda function is configured for function-level concurrent execution limit.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
false
cld-bpa-008
offensive
off_scanner_adjudication
boolean
aws_s3_bucket_public_access_block { block_public_acls=false ignore_public_acls=false block_public_policy=false restrict_public_buckets=false } and a bucket policy granting Principal "*" s3:GetObject — world-readable.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-106
offensive
off_scanner_adjudication
boolean
Terraform aws_lb_listener protocol="HTTPS" ssl_policy="ELBSecurityPolicy-TLS13-1-2-2021-06" plus aws_lb_listener_certificate; the ALB target group has a /healthz health check. TLS 1.2+ with health checks.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
false
cld-119
offensive
off_scanner_adjudication
boolean
Terraform aws_mq_broker "rabbit" { publicly_accessible = true } with default guest credentials still enabled. The message broker is internet-exposed.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cloud-verd-004
offensive
off_verdict
boolean
S3 bucket probe: GET https://company-data.s3.amazonaws.com/ returns HTTP 403 AccessDenied. HEAD request also returns 403. Bucket exists but is not publicly readable.
Is the probe result VULNERABLE?
{ "true": "the evidence confirms exploitability", "false": "the evidence is inconclusive, a false positive, or the control is effective" }
false
cld-285
offensive
off_scanner_adjudication
boolean
The GCP workload uses Secret Manager with secretmanager.versions.access and Workload Identity; no service-account key file is stored in the repository.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
false
cld-ck-009
offensive
off_scanner_adjudication
boolean
Checkov CKV_AWS_145 FAILED on aws_s3_bucket.pii: Ensure that S3 buckets are encrypted with KMS by default.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
true
cld-ck-011
offensive
off_scanner_adjudication
boolean
Checkov CKV_AWS_16 FAILED on aws_db_instance.customers: Ensure all data stored in the RDS is securely encrypted at rest.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
true
cld-ck-074
offensive
off_scanner_adjudication
boolean
Checkov CKV_AWS_129 FAILED on aws_db_instance.audit: Ensure that respective logs of Amazon RDS are enabled.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
false
cloud-seve-003
offensive
off_severity
boolean
AWS IAM user 'deploy-bot' has AdministratorAccess policy attached. Access keys were last rotated 847 days ago. User has no MFA enabled.
Is this finding HIGH severity or above?
{ "true": "HIGH or CRITICAL — significant data exposure, privilege escalation, or system compromise", "false": "MEDIUM or below — limited impact, hard to exploit, or requires preconditions" }
true
cld-ms-056
offensive
off_scanner_adjudication
boolean
cloudsploit PLUGIN rdsEncryptionEnabled: FAIL — RDS instance 'orders-db' is not encrypted at rest.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
true
cld-269
offensive
off_scanner_adjudication
boolean
Lambda function 'acme-processor' executes with an execution role that has the AdministratorAccess managed policy attached.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-025
offensive
off_scanner_adjudication
boolean
Terraform aws_ebs_volume "data": size=500, encrypted=false. A 500GB volume attached to a prod EC2 holds data unencrypted.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-213
offensive
off_scanner_adjudication
boolean
GCP firewall rule 'default-allow-ssh-world' permits tcp:22 ingress from source range 0.0.0.0/0 to all instances via target tags.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cloud-aler-005
defensive
def_alert_severity
boolean
GuardDuty HIGH: IAMUser/AnomalousBehavior — iam:CreateAccessKey called by user 'admin' for user 'backdoor-svc'. New access key created. No matching change request.
Is this alert high severity requiring immediate response?
{ "true": "active exploitation, data exfiltration, or privilege escalation in progress", "false": "low-priority, informational, or already mitigated" }
true
cld-ms-027
offensive
off_scanner_adjudication
boolean
trivy misconfig AVD-AWS-0089 (LOW): S3 bucket does not have server access logging enabled. aws_s3_bucket.assets.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
false
cld-bpa-007
offensive
off_scanner_adjudication
boolean
KMS key policy restricts kms:Decrypt to a single application role ARN; no wildcard principals.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
false
cld-302
offensive
off_scanner_adjudication
boolean
SNS topic 'acme-alerts' policy allows sns:Subscribe from Principal '*', letting anyone subscribe an endpoint and receive messages.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-311
offensive
off_scanner_adjudication
boolean
Azure APIM operation enforces a validate-jwt policy against the tenant authorization server before routing.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
false
cloud-cve_-008
offensive
off_cve_applicable
boolean
AWS RDS PostgreSQL 15.4. Checking CVE-2023-5868 (memory disclosure in aggregate function calls). RDS is managed — AWS patches are applied by the provider, but version 15.4 predates the fix (15.5).
Is this CVE applicable to this target?
{ "true": "the CVE applies and the target is vulnerable", "false": "the CVE does not apply — wrong version, config, or platform" }
true
cld-ck-046
offensive
off_scanner_adjudication
boolean
Checkov CKV_AWS_58 FAILED on aws_eks_cluster.main: Ensure EKS Cluster has Secrets Encryption enabled.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
true
cld-151
offensive
off_scanner_adjudication
boolean
aws_security_group ingress { from_port=6379 to_port=6379 cidr_blocks=["0.0.0.0/0"] } — Redis (no auth by default) exposed to the internet.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-059
offensive
off_scanner_adjudication
boolean
Terraform aws_iam_user_policy inline on "deploy": {"Effect":"Allow","Action":["iam:*","s3:*","ec2:*"],"Resource":"*"}. The CI user holds de-facto admin.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-092
offensive
off_scanner_adjudication
boolean
Terraform aws_security_group ingress referencing a prefix_list_ids = [aws_ec2_managed_prefix_list.corp.id] on port 22. SSH is allowed only from the managed corporate CIDR list.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
false
cld-307
offensive
off_scanner_adjudication
boolean
AWS API Gateway REST API 'acme-api' has an /admin resource with authorization type NONE and no API key or authorizer.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-ck-081
offensive
off_scanner_adjudication
boolean
Checkov CKV_AWS_51 FAILED on aws_ecr_repository.app: Ensure ECR Image Tags are immutable.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
false
cld-ms-064
offensive
off_scanner_adjudication
boolean
cloudsploit PLUGIN autoScalingGroupCooldown: WARN — Auto Scaling group 'web-asg' uses the default cooldown period.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
false
cld-063
offensive
off_scanner_adjudication
boolean
Terraform aws_security_group ingress { from_port=6379, to_port=6379, protocol="tcp", cidr_blocks=["0.0.0.0/0"] } on an ElastiCache Redis node with no AUTH. Open Redis is internet-reachable.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-ck-076
offensive
off_scanner_adjudication
boolean
Checkov CKV_AWS_353 FAILED on aws_db_instance.reporting: Ensure that RDS instances have performance insights enabled.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
false
cld-ms-063
offensive
off_scanner_adjudication
boolean
ScoutSuite: CloudFront distribution 'cdn' has no access-logging bucket configured (info).
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
false
cld-076
offensive
off_scanner_adjudication
boolean
Terraform aws_security_group "db" has only egress rules and no ingress block; nothing can connect inbound.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
false
cld-104
offensive
off_scanner_adjudication
boolean
Terraform aws_ecr_repository with image_tag_mutability="IMMUTABLE", scan_on_push=true, and an encryption_configuration { encryption_type="KMS" }; repository policy limited to the account. Private hardened registry.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
false
cloud-test-001
offensive
off_test_class
boolean
Target runs AWS Lambda with API Gateway (REST API, Node.js 20). Proposed test class: SSRF. Lambda functions can reach internal VPC resources and the metadata service.
Should this vulnerability test class run on this target?
{ "true": "the class is applicable to the detected technology", "false": "the class is impossible or irrelevant for this stack" }
true
cld-045
offensive
off_scanner_adjudication
boolean
Terraform aws_snapshot_create_volume_permission / aws_ebs_snapshot with create_volume_permission { group="all" }. An EBS snapshot of a prod volume is shared publicly.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-217
offensive
off_scanner_adjudication
boolean
AWS security group sg-1d4 allows inbound TCP 22 only from the corporate egress CIDR 203.0.113.0/24.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
false
cld-ms-057
offensive
off_scanner_adjudication
boolean
ScoutSuite: IAM policy 'deploy' allows 'sts:AssumeRole' on Resource '*' (danger).
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
true
cld-ck-062
offensive
off_scanner_adjudication
boolean
Checkov CKV_AWS_300 FAILED on aws_s3_bucket.multipart: Ensure S3 lifecycle configuration sets period for aborting failed uploads.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
false
cld-065
offensive
off_scanner_adjudication
boolean
Terraform aws_db_instance: publicly_accessible=true together with a security group allowing 5432 from 0.0.0.0/0. The Postgres DB is directly exploitable from the internet.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-262
offensive
off_scanner_adjudication
boolean
RDS instance 'acme-reports' has deletion protection on, is encrypted with SSE-KMS, and is deployed only in private subnets.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
false
cld-ms-039
offensive
off_scanner_adjudication
boolean
prowler check ecs_cluster_container_insights_enabled: FAIL — cluster batch has Container Insights disabled.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
false
cld-101
offensive
off_scanner_adjudication
boolean
Terraform aws_s3_bucket_policy allows Principal="*" s3:PutObject on a bucket wired as a CloudFront origin with no condition. Attackers can plant content served to all site users.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-248
offensive
off_scanner_adjudication
boolean
IAM role 'app-runtime' grants iam:PassRole scoped to a single role ARN with a condition iam:PassedToService = ecs-tasks.amazonaws.com.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
false
cloud-seve-004
offensive
off_severity
boolean
Azure Storage account exposes a SAS token in the application's client-side JavaScript. Token grants read access to a container holding user profile photos (public content).
Is this finding HIGH severity or above?
{ "true": "HIGH or CRITICAL — significant data exposure, privilege escalation, or system compromise", "false": "MEDIUM or below — limited impact, hard to exploit, or requires preconditions" }
false
cld-240
offensive
off_scanner_adjudication
boolean
IAM policy attached to the 'ci-deployer' user allows Action '*' on Resource '*' (full administrator via wildcard).
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
cld-132
offensive
off_scanner_adjudication
boolean
Terraform aws_iam_user "ci" has no inline or attached policy directly; permissions come only through an assumed role with a scoped policy and an ExternalId condition. No standing broad grant.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
false
cld-ms-028
offensive
off_scanner_adjudication
boolean
prowler check ec2_securitygroup_allow_ingress_from_internet_to_port_22: FAIL — security group sg-0a1 allows 0.0.0.0/0 to port 22.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
true
cld-ck-055
offensive
off_scanner_adjudication
boolean
Checkov CKV_AWS_117 FAILED on aws_lambda_function.handler: Ensure that AWS Lambda function is configured inside a VPC.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
false
cld-ms-045
offensive
off_scanner_adjudication
boolean
Snyk SNYK-CC-00312 (High): ECR repository allows public pull through its repository policy.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / best-practice nit / not reportable alone" }
true
cld-130
offensive
off_scanner_adjudication
boolean
Terraform aws_s3_bucket_policy using NotPrincipal is avoided; instead an explicit Allow to one role ARN and a Deny for aws:SecureTransport=false. Clear allow-list plus TLS enforcement.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
false
cld-303
offensive
off_scanner_adjudication
boolean
GCP Pub/Sub topic 'acme-events' IAM grants roles/pubsub.subscriber to allUsers.
Is this a real, reportable security finding?
{ "true": "a real, exploitable misconfiguration or vulnerability", "false": "informational / safe / hardened / not reportable alone" }
true
End of preview.

Haruspex security-decision benchmark (in-distribution splits)

Held-out test splits for haruspex. Each .jsonl row is {type, state, question, criteria/candidates/levels, target}. These are in-distribution splits of the training data; independent results are reported in the model card against public datasets (e.g. deepset/prompt-injections, the OWASP Benchmark).

  • data/security_test.jsonl — web/offensive security decisions (own synthetic)
  • data/mobile_test.jsonl — mobile app security (own synthetic)
  • data/cloud_test.jsonl — cloud posture (own synthetic)
  • data/network_test.jsonl — network security (own synthetic)
  • data/guide_test.jsonl — Microsoft GUIDE incident triage (CDLA-Permissive-2.0 (Microsoft GUIDE))
Downloads last month
51