Security-first scanning
Surface vulnerabilities and risky patterns before they become incidents.
VELLUM
CLI Agent for healthier codebases
vellum helps engineers and hobbyists continuously improve code quality so projects stay secure, maintainable, and scalable as they grow.
Two execution modes
Mode 1: Direct CLI invocation
Run on demand when you want targeted analysis and refactor suggestions before shipping.
Mode 2: Post-commit action
Trigger automatically in the background after each commit to keep quality checks continuous.
Introduction
Your problem
Your agents can leak credentials, require repetitive refactor prompts, and ignore their configuration files. Those failures compound between commits.
What vellum does for you
vellum is the agent that keeps your other agents in check. It uses best-in-class SAST and code hygiene tooling, either on-demand or automatically post-commit.
vellum terminal
$ vellum-code
Vellum Code
Analyze and refactor your code for security, maintainability, and scalability
Select analysis mode:
› 📝 Git Diff - Analyze files from latest commit
📁 Files/Directories - Select files or directories to analyze
⠋ Analyzing 7 file(s)...
Current step: Running SonarQube analysis...
• SonarQube security scan
• Code quality assessment
• LLM-powered refactoring suggestions
Features
Surface vulnerabilities and risky patterns before they become incidents.
Get concrete refactors that improve readability, structure, and adaptability.
Enforce best practices using SonarQube quality gates to achieve the highest standards.
Use commit-level context to focus feedback on what actually changed.
Review suggestions and apply only what matches your standards.
Useful whether you maintain production systems or side projects.
How It Works
Step 1
Choose run mode
Step 2
Analyze changed files
Step 3
Generate clear fixes
Step 4
Apply what you approve