Documentation
Set up Cefense in this repo: cefense.com/skill.mdLooking for something specific? Press to search every page.
Documentation
How Cefense turns a reproduced attack into a verified repair inside your repository.
CLI
Connect a repository, scan it, read what came back, and ship the fix from your terminal.
Agent integration
Give Claude Code, Cursor, Codex, and the rest the ability to close a vulnerability.
Explore by capability
What a scan reads, what it produces, and what it never touches.
Why a finding is validated before it reaches you, and what confidence means.
The smallest safe repair, written as a diff you can read.
From a proposed patch to a merged branch, without bypassing anything.
GitHub, GitLab, and Bitbucket, and what each one grants.
What Cefense reads, what is stored, and how to delete all of it.
skill.md
A single page written for an agent that has never seen Cefense. It installs the CLI, signs the user in, connects the repository, and finishes by installing the full skill.
Read skill.mdResearch and field notes
Long-form writing on frontier AI, verification, and security: the reasoning behind how Cefense closes an attack path.
Read the blog