Trident never stores long-lived credentials. It assumes the read-only role
on each scan cycle using short-lived tokens, so rotating or revoking the
role immediately stops all access.
Connect an AWS account
AWS is the most common first connection. Trident provides a CloudFormation template to create the role in one click, or you can apply the IAM policy JSON manually.Open the Add Account flow
In the Trident dashboard, navigate to
Cloud → Add Account → AWS.
Copy your External ID
Trident displays a CloudFormation template link and an IAM policy JSON
document. Before you leave this page, copy your External ID — you
will paste it into the IAM trust policy in the next step. The External ID
is unique to your Trident project and prevents confused-deputy attacks.
Create the IAM role in your AWS account
In your AWS account, create a new IAM role with the trust policy Trident
provides. The trust policy allows Trident’s scanner principal to assume
the role, scoped to your External ID. Attach the Trident-provided
read-only permission policy — it covers compute, IAM, networking,
storage, and secrets services. No write permissions are included.If you prefer one-click setup, launch the Trident-provided
CloudFormation template instead — it creates the role and attaches the
policy automatically.
Paste the role ARN and connect
Back in the Trident dashboard, paste the ARN of the role you just
created (format:
arn:aws:iam::<account-id>:role/<role-name>) and
click Connect.Connect Azure, GCP, or Kubernetes
Each provider follows the same pattern — you create a read-only credential in your environment and hand Trident a reference to it.- Azure
- GCP
- Kubernetes
Navigate to Cloud → Add Account → Azure. Trident walks you through
creating an Azure service principal with a Reader role assignment
scoped to your subscription or management group. You provide the
tenant ID, client ID, and client secret. Trident encrypts the credential
at rest using AES-256-GCM and uses it only during scan cycles.
What permissions Trident requests
Trident requests the minimum read-only permissions required to build the security graph. The scope covers:| Surface | Examples |
|---|---|
| Compute | List/describe EC2 instances, Lambda functions, EKS clusters, pods |
| IAM | List roles, policies, users, groups, and effective permission simulation |
| Networking | VPCs, subnets, security groups, load balancers, DNS |
| Storage | S3 bucket metadata, ACLs, and policies (not object contents) |
| Secrets | Secrets Manager / KMS key metadata (not secret values) |
| Databases | RDS, DynamoDB, Cloud SQL instance metadata |
Ingest output from existing scanners
If you already run Prowler, Trivy, Kubescape, Falco, CloudQuery, Steampipe, TruffleHog, or similar tools, you can push their JSON output directly into Trident’s security graph. This enriches your attack path analysis with data from tools you already trust. Send aPOST request to the scanner ingest endpoint using your project’s API key for Basic auth:
scanner values is:
| Value | Tool |
|---|---|
falco | Falco / falcosidekick runtime alerts |
trivy | Trivy container and repository scans |
kubescape | Kubescape Kubernetes posture |
cloudquery | CloudQuery asset and edge inventory |
steampipe | Steampipe cloud inventory |
threatmapper | ThreatMapper cloud and container topology |
pmapper | PMapper effective-permission graph |
neuvector | NeuVector deep runtime security |
stackrox | StackRox / RHACS Kubernetes posture |
ciem | Generic CIEM output |
gcp_ciem | GCP-native CIEM output |
azure_ciem | Azure-native CIEM output |
terraform | Terraform plan static analysis |
checkov | Checkov IaC scanning |
tetragon | Tetragon eBPF runtime events |
trufflehog | TruffleHog secret detection |
osv | OSV vulnerability feed |
cloudsplaining | Cloudsplaining IAM analysis |
dspm_content | DSPM content-tier data classification |
mcpSafety | MCP tool-call safety scanner |
scanId and a 202 Accepted status — findings appear in your inbox once Trident processes the job.