
VISaR
Vulnerability Identification, Scanning & Reporting
A free, open-source command-line tool that scans GitHub repositories against known vulnerability databases - built for data platform teams who need a repeatable evidence record before approving open-source dependencies.

One file. Every scan. No server, no setup. VISaR v1.1.0 rolls every scan in your data/ directory into a single self-contained HTML dashboard. Filter by severity, switch between scans, and expand any row for full detail - all in one file you can email or share.
The problem
Most of your codebase is code you didn’t write.
“VISaR gave us an evidence trail that cut approval times for new technologies from weeks to days.”
Every open-source library you pull in, along with every dependency it brings with it, is a potential entry point into your data platform. In regulated environments, “we think it’s fine” isn’t an answer an auditor accepts, and AI assistants now generate dependency choices faster than anyone can vet them by hand.
VISaR turns that vetting into a single, repeatable command that takes a minute or two. Point it at a repository and get back a structured, severity-rated report you can keep as evidence. Every dependency approval is backed by the same check, run the same way, every time.
It started when a client asked how to quantify their open-source risk. We built VISaR for our own client work, and we run it on every platform we deliver.
How it works
From a repo URL to an evidence record - in four steps.
- 01
Point it at a repository
Give VISaR any public GitHub URL or a text file of them for batch mode. OSSF Scorecard scans the project and surfaces its known vulnerabilities.
- 02
Findings are enriched via OSV
Each vulnerability ID is sent to the OSV database to pull back severity ratings, plain-text descriptions, and aliases, so a raw ID becomes a decision you can act on.
- 03
A structured report is written
Results are compiled into a clean CSV or JSON file, named by date and repository. Open it in Excel, diff it in version control, or feed it into your own pipeline.
- 04
Consolidate into one dashboard
Roll every scan into a single self-contained HTML dashboard. Filter by severity, switch between scans, expand rows for full detail, then email it or drop it in a shared drive.
Under the hood, VISaR leans on best-in-class open source: the OSSF Scorecard for identification and the OSV database for enrichment. Output is available as CSV, JSON, or the interactive HTML dashboard above.
Who it’s for
Built for teams that have to prove what they scanned.
Data Engineers & Platform Teams
Evaluate open-source libraries before they enter your data stack. Keep VISaR's structured output as your evidence record, version-controllable and auditor-ready.
Software Engineers
Validate your own codebase before a release. Confirm your dependency graph carries no known vulnerabilities into the milestone you're shipping.
Independent Developers
Got an AI-generated dependency you're not sure about? Run it through VISaR first, a fast, objective check before it lands in your codebase.
Quick start
Up and scanning in under 5 minutes.
Clone, sync dependencies with uv, complete a one-time setup, and point VISaR at any public GitHub repository.
- Python 3.12+ (managed automatically by uv)
- Docker Desktop (2 GB available memory)
- GitHub personal access token with
public_reposcope - Network access to
api.github.comandapi.osv.dev
Current limitations
- Public GitHub repositories only (private repos are not supported yet).
- Batch scans run sequentially (one repository at a time).
- Docker Desktop must be running before a scan.
# 1. Clone and syncgit clone https://github.com/AtLongLastAnalytics/visarcd visaruv sync# 2. One-time setupcp .env.example .env# set VISAR_AUTH_TOKEN in .envdocker pull gcr.io/openssf/scorecard:stable# 3. Scan a single repository (writes CSV to data/)uv run visar https://github.com/matplotlib/matplotlib# 4. Batch scan from a listuv run visar --batch repos.txt# 5. Build the interactive dashboarduv run visar-dashboard # writes data/visar_dashboard.html# Want to look first? Preview with the bundled example data:uv run visar-dashboard examples
FAQ
Common questions.
How is VISaR different from Dependabot or Snyk?
Can we use VISaR commercially?
What does a customization engagement look like?
Do we own the custom features we pay for?
Do you offer support contracts for the free open-source version?
Private deployments and custom features.
We deploy VISaR inside your security boundary and extend it for your stack and compliance needs.
VISaR v1.1.0: Batch Scanning and Interactive Dashboard
VISaR v1.1.0 adds batch scanning across multiple repos and a self-contained HTML dashboard for sharing results. No server, no setup, just a file your team can open and explore.
VISaR: A Code Scanning Tool for Data Platform Engineers
VISaR is a free, open-source tool for scanning GitHub repositories against known vulnerability databases. Built after a client asked how to quantify their open-source risk.