Authentication

Last updated: Oct 28, 2026


When you authenticate a scan, the Fluid Attacks platform records who did each run. How you authenticate depends on who runs the scan: a person at a workstation signs in as themselves, and a pipeline or container authenticates as your group.

Start here

Who runs the scanAuthenticates asHow
You, on your own machineYouBrowser login
A pipeline or containerYour groupA Group token, supplied one of the two ways below

If you configure CI/CD, use a Group token. If you run a scanner on your laptop, sign in through your browser. You do not have to use a token.

Authenticating as your group

A Group token is a credential for your group only. The command-line scanners of Fluid Attacks and the CI Gate use it. You can give it in two ways.

Way to supply itBest forWhat you keep
Short-lived, via OpenID ConnectAny system that can mint OIDC tokensNothing — minted per job
Long-lived secretSystems that cannot mint OIDC tokensThe Group token, as a secret

Your CI/CD provider mints a short-lived, signed token for each job. Your scan names its group with --group. Fluid Attacks then makes sure that the token agrees with a trust that you define one time, and gives your group access for that run. You do not have to keep or rotate a secret.

Follow the OIDC federation guide to set it up.

Long-lived secret

If OIDC is not available, generate a long-lived Group token and store it as a secret. Generate it with the DevSecOps → Manage token flow of your group (read CI Gate configuration). Then, give it to the scan as the INTEGRATES_API_TOKEN environment variable:

export INTEGRATES_API_TOKEN="<your-group-token>"
ss scan --strict .

The scan uses the token only to identify the caller. The scan does not show the token or write it to logs.

Which should I use?

  • Running a scan yourself, on your own machine? Sign in through your browser. Nothing to store, and the run is attributed to you.
  • Can your CI/CD system or runtime mint OpenID Connect tokens? GitHub Actions, GitLab CI, Azure DevOps, Kubernetes, and many others can. Use OIDC, because you do not have to keep a long-lived secret.
  • If it cannot, keep a long-lived Group token as a secret.

The two group options authenticate the run in the same way. The only difference is if you keep a long-lived secret or not. A browser login is different: it authenticates you, not your group. If you also give a Group token, the scan uses the Group token.

A separate personal API token authenticates you to the GraphQL API, not a scan — see API basics.

What authenticating gets you

For the SAST, SCA, and secrets scanning CLIs, a run that resolves to a group also gets two policies of that group. They are the exclusions and the severity-breaking policy on the Fluid Attacks platform. The run applies them automatically:

  • Exclusions: The excluded subpaths of the Git root replace the exclude list of your configuration file. The scan does not put the two lists together. If the root has no excluded subpaths on the Fluid Attacks platform, the run scans all paths, and your local exclude list has no effect.
  • Severity threshold: The minimum CVSS score to break the build of your group becomes the threshold of the run. If you also set a strictness_threshold, the lower of the two values applies. The threshold applies only when you set the run to break: with --strict, strict: true, or a strictness_threshold. With only strict, the run fails only on vulnerabilities at or above that score. If the group has no minimum CVSS score, the run fails on each vulnerability.

For more information, read Platform policy sync.

On this page