FlowlytFlowlyt

CI/CDiswheresecretsgotoleak.

Flowlyt scans GitHub Actions workflows, GitLab CI pipelines, and other CI/CD configurations for exposed credentials, OIDC misconfigurations, and unvalidated supply chain dependencies — catching what code review misses.

.github/workflows/deploy.yml5 findings
1name: PR Validation
2
3on:
4 pull_request_target:CRIT FLW-014
5 types: [opened, synchronize]
6
7permissions: write-all
8
9jobs:
10 review:
11 runs-on: ubuntu-latest
12 steps:
13 - uses: actions/checkout@v4FAIL FLW-001
14 with:
15 ref: $${{ github.event.pull_request.head.sha }}CRIT FLW-021
16 persist-credentials: trueWARN FLW-022
17 - name: Lint PR title
18 run: echo "$${{ github.event.pull_request.title }}" | bashCRIT FLW-031
19 - uses: reviewdog/action-setup@mainFAIL FLW-011
2 critical2 failures1 warning

Leaked secrets in CI

API tokens, database credentials, and signing keys committed directly into workflow files or passed as unmasked environment variables. They persist in git history long after the secret is rotated.

One exposed NPM token shut down a company's deployment pipeline for three days.

Over-permissioned GitHub tokens

OIDC configurations granting write access to all tokens when only a single scope is required. Misconfigured workflow permissions that allow any job to push code or modify releases.

One misconfigured id-token scope gave an external contributor write access to the main branch.

Unvalidated third-party Actions

Actions referenced by mutable tags like @v3 or @main instead of immutable SHA commits. A compromised upstream action runs with the same permissions as your workflow — silently.

One exposed token ended a company's seed round.

Step 01

Connect your GitHub org or GitLab group.

Authorize Flowlyt with a single OAuth prompt. It requests the minimum permissions required — read access to workflow files and the ability to post PR/MR review comments.

$ flowlyt auth --org acme-corp --provider github
Authenticating with GitHub...
✓ OAuth token verified
✓ Repository access: 24 repos
✓ Webhook configured for push events
Ready. Run `flowlyt scan` to begin.

Step 02

Flowlyt scans every push, PR/MR, and scheduled run.

Every workflow trigger — push, pull_request, merge_request, schedule — is analyzed automatically across GitHub Actions, GitLab CI, and other supported platforms.

$ flowlyt scan --workflow .github/workflows/deploy.yml
Scanning deploy.yml...
FLW-001 FAIL actions/checkout@v3 (line 14)
↳ Action not pinned to SHA commit
FLW-003 WARN id-token: write (line 11)
↳ OIDC permission scope overly broad
FLW-005 CRITICAL DEPLOY_KEY hardcoded (line 19)
↳ Secret detected in plaintext env var
3 findings (1 critical, 1 failure, 1 warning)

Step 03

Findings land as inline PR comments and a SARIF report.

Each finding is posted directly on the offending line of the pull request. The full report exports as SARIF 2.1, compatible with GitHub Code Scanning and any SAST platform.

// flowlyt[bot] · deploy.yml, line 19
//
// [CRITICAL] FLW-005: Hardcoded secret detected
//
// Environment variable DEPLOY_KEY contains what
// appears to be a production API key in plaintext.
//
// Resolution: use a GitHub Actions secret instead:
// DEPLOY_KEY: ${{ secrets.DEPLOY_KEY }}
//
// Rule reference: https://flowlyt.io/rules/FLW-005

Intelligent agentic analysis, not noisy matching.

Flowlyt uses an intelligent agentic system to reason about workflow intent, data flow, and execution behavior. That context helps reduce false positives and makes the findings more trustworthy for engineering and security teams.

01

Context-aware analysis

Flowlyt reasons about triggers, permissions, job boundaries, and execution context instead of flagging every risky-looking pattern in isolation.

02

Lower false positives

The system filters noisy matches by understanding which findings are actually reachable and meaningful in the real workflow path.

03

Actionable findings

Security teams get fewer dead-end alerts and more precise findings they can trust, prioritize, and remediate quickly.

Secrets Detection

Scans workflow files and environment variable blocks for hardcoded tokens, API keys, and credentials — including entropy-based detection for unrecognized formats.

OIDC Misconfiguration Analysis

Detects overly permissive id-token scopes and flags non-minimal permission sets that violate the principle of least privilege in GitHub Actions OIDC configurations.

Action Pinning Enforcement

Enforces SHA-commit pinning on all third-party Actions, blocking mutable tag references like @v3 or @main that expose workflows to upstream supply chain attacks.

Multi-Ecosystem Coverage

Analyzes GitHub Actions, GitLab CI, and other CI/CD ecosystems with platform-aware rules for risky permissions, insecure triggers, and unsafe pipeline behavior.

SARIF Output

Exports findings in SARIF 2.1 format, compatible with GitHub Code Scanning, SonarQube, Semgrep, and any SAST platform that accepts standard static analysis results.

GitHub PR Annotation

Posts inline review comments on the exact line of each violation, with remediation guidance, rule reference, and severity classification — directly in the pull request.

We found a leaked NPM token that had been live in CI for eleven months. Flowlyt caught it in the first scan. We had no idea it was there.

— Platform Lead, Series B fintech

The SARIF output plugs directly into our existing security dashboard. Zero configuration on our end. It just appeared in Code Scanning like any other tool.

— Staff Engineer, enterprise SaaS (2,000 employees)

Every third-party Action in our org is now SHA-pinned. Enforcing that workflow alone was worth the annual contract — it takes that work completely off our plate.

— DevSecOps Lead, Series C infrastructure startup

Join the waitlist.

Flowlyt is in closed preview. Drop your email — we'll reach out when your team's slot opens, with full access and a setup session included. No credit card required.

or schedule a demo

01

Full platform access

Unlimited repositories, 90-day history, and all detection rules from day one.

02

Dedicated setup session

A 30-minute call with the team to connect your org and review your first scan report.

03

Direct feedback loop

Your security findings shape what we build next. Early teams set the product roadmap.