LaunchKit · 2026
Back to Skills

ralph-security

Comprehensive security audit with 100 iterations (~30-60 min)

0
401 downloads
by @dorukardahan

Setup & Installation

openclaw skills install @dorukardahan/ralph-security

Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:

npx clawhub install ralph-security

Version History

v3.0.0Feb 19, 2026 - Version note by skill developer:

Latest update

What This Skill Does

Runs a structured 100-iteration security audit covering OWASP Top 10, authentication, secrets detection, infrastructure, and code quality. Each iteration performs one focused check with a pass/fail/warn result, confidence rating, and fix recommendation. Progress is saved to a report file every 10 iterations.

Combines reconnaissance, OWASP coverage, secret scanning, and infrastructure review in a single structured run rather than requiring separate tools for each concern.

When to use it

  • Weekly security check before a production deployment
  • Onboarding a new open-source project you're adopting
  • Pre-release audit of an API with authentication endpoints
  • Reviewing a containerized app for Docker and network misconfigurations
  • Scanning a codebase for accidentally committed secrets or API keys

Example Workflow

Here's how your AI assistant might use this skill in practice.

INPUT

User asks: 'Run a security audit on this project'

AGENT
  1. 1Detects stack and infrastructure by reading package.json, Dockerfile, and CI config files
  2. 2Enumerates public endpoints, auth requirements, and environment variables in Phase 1
  3. 3Steps through OWASP Top 10 checks in Phase 2, reading actual code before marking any finding as FAIL
  4. 4Audits JWT usage, OAuth flows, secret detection, and rate limiting in Phase 3
  5. 5Checks container config, network exposure, TLS, SSH, and database permissions in Phases 4-5
  6. 6Generates a final scorecard and saves the full report to .ralph-report.md
OUTPUT

A structured .ralph-report.md with per-finding severity, confidence level, and fix recommendations across all 100 checks