Security checks that run before you ship

Scanners, automation and CI gates, with output in formats your tools already read.

Tokens pinned, gates quiet

Designed against NIST SSDF, SLSA and OWASP Top 10 CI/CD Security Risks.

  1. TriggerLeast-privilege tokensScoped to one job and one repo, short-lived.
  2. ActionsPinned actions and toolsPinned to a commit or a checksum. Moving tags are out.
  3. SecretsSecrets stay secretMasked, scoped and never echoed to a log.
  4. ScannersSmall tools, easy to auditDependency-light CLIs that do one job on your own machines.
  5. FindingsNoise under controlReport-only first, then gating. A check that cries wolf gets tuned or cut.
  6. ReportStandard outputSARIF 2.1.0 and JSON.

Gates, scanners and rules

  • CI/CD security gates

    Checks in your CI that fail a build on what you decide matters, and stay quiet about the rest.

  • SAST, DAST and SCA pipelines

    Static, dynamic and dependency checks, tuned so nobody gets paged for noise.

  • Custom rules

    Semgrep and CodeQL rules for the bug patterns that keep turning up in your code.

  • Command-line scanners

    Python, Go or Swift CLIs for what off-the-shelf tools miss, with text, JSON and SARIF output.

  • Reporting

    SARIF 2.1.0 into code scanning, and a summary people actually read.

  • Recon and inventory

    Asset and subdomain inventories for scopes you’re authorised to test, kept current.

A pipeline that holds the keys to production is a target too. We build it like one.

Where your code already builds

  • Languages

    PythonGoSwift

    Python for glue, Go for fast single-binary scanners, Swift for macOS internals.

  • Pipelines

    GitHub ActionsGitLab CI

    Gates live where your code already builds.

  • Analysis

    SASTDASTSCA

    Static analysis on every change, dynamic scans on staging, dependency checks on every lockfile.

  • Output

    SARIF

    SARIF 2.1.0, so findings land in the tools you already use.

Tools we use: Semgrep, CodeQL, ZAP, Trivy and Gitleaks.

Will this slow our builds down?

We measure it. Fast checks run on every pull request, slow scans nightly or before a release. Gates start in report-only mode.

Can the tools run on our own infrastructure?

Yes. We favour tools that run on your own runners and send nothing out.

We run it on ourselves

  • vulnaratechnologies.com

    • A post-build check fails the build on any inline script or style, so the strict CSP can never be switched off by accident.

Tell us what should fail the build.

Where your code builds, and what keeps slipping through. We’ll quote it in writing.