58 checks · fixes as pull requests · GitHub + GitLab

Every scanner finds problems. This one opens the pull request.

Leaked secrets, known CVEs in dependencies, unsafe GitHub Actions, Terraform and Kubernetes misconfiguration. Where a fix can be written, it arrives as a pull request you review and merge — not a ticket you have to write.

58
checks
~5s
per scan
9
categories
AI
recommendations
// problem

Most repositories are not production-ready

Hardcoded secrets, no CI/CD, outdated dependencies — teams find out only when something breaks.

🔓
Secrets in code
A committed API key stays valid until somebody notices it, and git history keeps it long after the file is deleted.
🚫
No automation
Without CI, every deploy is a manual step someone can skip, and every test run depends on somebody remembering to do it.
⚠️
Vulnerable dependencies
Outdated packages without lock files mean unpredictable builds and known CVEs in production.
// what we check

Up to 58 checks — only the ones that fit your project apply

A library is not asked for a Dockerfile, a docs site not for tests. About 5 seconds per public repository.

01 / SUPPLY CHAIN
⛓
CI/CD supply chain
Actions pinned to a commit SHA, least-privilege workflow tokens, unsafe pull_request_target, script injection, self-hosted runners, GitLab CI includes.
02 / SECRETS
🔐
Secrets
Keys and tokens anywhere in the tree and in git history; for connected repositories, whether a found key still works.
03 / CVE
🛡
Known vulnerabilities
Every dependency checked against OSV.dev, with whether the vulnerable package is actually imported.
04 / IAC
☁
Terraform and Kubernetes
Remote state and locking, 0.0.0.0/0 ingress, hardcoded credentials, pinned providers; pod security, probes and limits.
05 / CONTAINERS
🐳
Containers
Non-root user, pinned base images, multi-stage builds, healthchecks, resource limits — in every Dockerfile, not just the root one.
06 / HYGIENE
📚
Repository hygiene
README, license, security policy, changelog, CODEOWNERS — shown for completeness, kept out of the score.
Dockerfile
.gitignore
CI/CD pipeline
Secrets in code
Healthcheck
README
Lock file
Branch protection
Non-root Docker
Dependabot
.env.example
CHANGELOG
License
PR templates
Tests in CI
// how it works

Three steps to the pull request

No agents or manual configuration

01
Sign up
Email and password, or with Google, GitHub or GitLab. No card, no OAuth app to set up.
free
02
Connect GitHub or GitLab
Sign in with GitHub or GitLab, or paste an access token. An audit only reads; nothing is written to a repository until you open a fix yourself. Self-hosted GitLab on internal networks connects with a token.
read to audit · write to fix
03
Get the report — and the fix
A 0–100 score, a breakdown by category, an AI analysis, and a one-click pull request for the findings that have a fix.
~5 seconds
// supported platforms

Works with your stack

GitHub
GitLab
🐳
Docker
🐍
Python
🟨
Node.js
🐹
Go
// see it work

Run it on a repository you know

Any public repository, no account needed. Here are a few to start with.

Scan any repository free
// pricing

Start free, pay when it pays for itself

Annual billing saves 20%. Team seats are included — you pay per organisation, not per engineer.

FREE
Free
$0 / forever
Personal projects and a real look at the product

Unlimited public repository scans
3 private repositories
1 AI report + 1 PDF per month
Score badge for your README
Score history and trend
Scan a repository
TEAM
Team
$99 / month
or $79/mo billed yearly — $950
One organisation, seats included

Everything in Pro
Organisation dashboard
Weekly trend across all repositories
GitLab self-hosted
Start with Team →
COMPLIANCE
Compliance
$299 / month
or $239/mo billed yearly — $2870
For the audit you already have a date for

Everything in Team
SOC 2 and ISO 27001 control mapping
Evidence export for auditors
SBOM export (CycloneDX, SPDX)
Full audit history retention
Start with Compliance →

Questions engineers ask first

Is it safe to give DevOpsCheck access to my repositories?

Sign in with GitHub or GitLab, or paste an access token. An audit only reads, and nothing is written to a repository until you open a fix yourself. Note that GitHub sign-in grants the repo permission, which GitHub does not split into read and write — for strictly read-only access, paste a fine-grained token with read permissions instead. Stored tokens are encrypted. The Trust page lists what is read and what never leaves. Trust page →

What is sent to an AI provider?

Nothing, unless AI analysis is switched on in your profile. When it is: the failed checks with our recommendations, facts about the repository, and up to 12 configuration files (Dockerfiles, CI workflows, Kubernetes, Terraform) — never application source code, and anything that looks like a credential is masked first. Turn it off and reports keep their advice, written by built-in rules.

How is it different from Dependabot or OpenSSF Scorecard?

Dependabot updates dependencies; Scorecard grades security practices. DevOpsCheck covers both areas plus Docker, Terraform, Kubernetes and GitHub Actions supply chain in one report with one score — and where a fix can be written, it opens the pull request.

Does it work with self-hosted GitLab?

Yes. Connect with a GitLab token, including instances on internal networks with their own certificates and DNS. Fixes arrive as merge requests.

Can it run on-premises?

Not as a packaged install yet. If you need it inside a closed network, write to us — it is on the roadmap.

How much does it cost?

Scanning public repositories is free and needs no account. A free account covers private repositories up to a limit; paid plans add AI analysis, fix pull requests and team features.

Point it at a repository you care about

Public repositories are free and need no account.

Scan a repository → Create an account