The security reviewer that lives in your editor.
The Jikida Skill is a set of security instructions your AI assistant follows while it writes code, so it guards what you build and runs a real scan, pentest or repo-scan on demand, without leaving Claude Code, Cursor, Windsurf or Copilot.
- One command to install
- Guards code as it is written
- Scan, pentest and repo-scan on demand
- Works with 9 AI assistants
Install
One command and your AI knows security.
No account, no token to paste. The installer detects which assistants you use and writes the Skill for each, plus a /jikida-scan command and a hardened .gitignore.
Then type /jikida-scan in your editor, or just say “scan this for security issues”. Your assistant reads the Skill and checks your code the way a reviewer would, before you ship.
What the Skill does
Protection applied as you build, not bolted on after.
A Skill is knowledge your assistant carries, not a tool you call. Once installed, it writes safer code by default and catches the classic mistakes the moment they appear.
What it can run
Ask in plain English. It runs the real thing.
The Skill is wired to Jikida’s tools, so your assistant runs the same checks the dashboard runs and reads them back inline, it does not guess.
Works with your assistants
One install, every editor.
The installer writes the right file for each assistant it finds: Claude Code, Cursor, Windsurf, GitHub Copilot, Codex, Gemini, Cline, Aider and Continue. No per-editor setup.
The Skill teaches your AI to think about security and reach for the right tool. The MCP gives it those tools to run. The SDK puts the protection in your running app. One command sets up all three, you do not have to choose.
Frequently asked questions.
Everything people ask before they start. Still unsure? We are one message away.
A Skill is a packaged set of instructions your AI assistant loads and follows. The Jikida Skill teaches Claude Code, Cursor, Windsurf and others how to write secure code and when to run a scan, a pentest or a repo secret-scan.
They pair. The Skill is the knowledge (how to think about security and which tool to reach for); the MCP server is the set of tools it runs. See the MCP server →
Claude Code, Cursor, Windsurf, GitHub Copilot, Codex, Gemini, Cline, Aider and Continue. One command writes the right file for each one it detects.
No. Installing the Skill is a local command that writes files into your project. The free scan and header tools run without a key; deeper scans and pentests use your Jikida account when you have one.
Yes. The Skill is wired to Jikida’s real tools, so “run a pentest on staging” runs the same active, OWASP and CWE-mapped pentest the dashboard runs.
Give your AI a security reviewer.
One command installs the Skill for every assistant you use. Start writing safer code this session.