Code security scanner features From the source to the sink, with the fix.

scancodebase follows untrusted data through your code, checks what your code depends on and reads your configuration. Here is what that covers.

Features

Code analysis (SAST)

Untrusted data is followed from where it enters to where it does damage, across functions and files, with models of the usual frameworks such as Spring, Django, Flask, Express, Laravel, Rails and ASP.NET.

Dependencies (SCA)

Every package and lock file is matched against public vulnerability sources. You get the version that fixes each issue and which update removes the most risk.

Secrets

Hundreds of kinds of credentials: cloud, CI/CD, version control, package registries, messaging, payments, databases and private keys. Validated offline and always shown masked.

Infrastructure as code

Kubernetes, Helm, Dockerfile, Terraform and CloudFormation, normalised to what the cluster or the cloud will really apply. Covers AWS, GCP and Azure with CIS and NSA references.

Pull requests

Only what is new in each pull request, with a fingerprint that tolerates code being moved. Findings arrive as comments and as SARIF.

Your own rules

Write rules for your codebase with a syntax compatible with Semgrep. Suppress a finding with a reason, an owner and an expiry date.

Compliance mapping

Each finding mapped to CWE, OWASP Top 10, OWASP ASVS and PCI DSS.

Reports and exports

A plain-language explanation of every finding. Exports in SARIF for GitHub, GitLab and Azure DevOps, plus JSON, HTML and PDF.

In detail

Go deeper into each one

Coverage

What it understands

Web and services
Python, JavaScript, TypeScript, Java, Kotlin, Scala, Groovy, Go, C#, VB.NET, PHP and Ruby.
Systems
C, C++ and Rust.
Scripts and others
Bash, PowerShell, Perl, Lua, Solidity, Elixir, Erlang, Apex and R.
What we catch
Injection (SQL, command, code, template, LDAP, XPath), XSS, SSRF, path traversal, open redirect, unsafe deserialization, weak cryptography, insecure cookies, unverified TLS and memory errors in C and C++.

What makes it different

Why scancodebase is different

A trace for every finding

You see where the untrusted data comes in and every step it takes to the risky call, across functions and files.

A proposed fix, not just a warning

Each finding comes with a safer way to write that line, for its language and its kind of flaw.

Priorities, not endless lists

Findings are ranked by whether they can be reached from an entry point, so the report starts with what matters.

Only what changed

On a pull request we report what is new, even if the code has moved, so old findings do not bury the new ones.

Secrets never leave

A key we find is shown masked and is checked offline. It is never tested against the provider it belongs to.

Nothing goes to third parties

Your code is not sent to an outside service. Results do not depend on an AI model, so the same code always gives the same report.

On the way

What we are building next

Coming soon

Cloud posture

Attack paths from the internet to your data across AWS, GCP and Azure, joined with your code and images.

Coming soon

AI security

Hidden instructions in prompts and risky agent and MCP configuration.

Coming soon

Mobile apps

Static analysis for mobile applications.

Coming soon

Self-hosted and air-gapped

The same product inside your own infrastructure, with no internet connection.

Same team, same short reports

Scan your Docker images too

scancontainer checks the image you build from this code: vulnerabilities, malware, tampered files and secrets.

Go to scancontainer

Scan yours before the next merge.