LLM Pentest Juice Shop — Security skill for Claude Code
Demonstration of LLM coding agents performing automated black-box pentesting and white-box code review against OWASP Juice Shop, with consolidated security reporting.
How to install LLM Pentest Juice Shop
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open csautter/llm-pentest-juice-shop and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What LLM Pentest Juice Shop does
Demonstration of LLM coding agents performing automated black-box pentesting and white-box code review against OWASP Juice Shop, with consolidated security reporting.
Alternatives in Security
- Skills Audit Report — Date: 2026-02-15 Auditor: Automated Skill Quality Audit Scope: Recently added skills in business-growth/, fina 5.3k ★
- Promptmap — Security scanner for custom LLM apps 1.2k ★
- Clawsec — Security skill suite with drift detection, automated audits, and skill integrity verification 798 ★
README
OWASP Juice Shop — LLM Security Demo Environment
A reproducible dev-container environment that demonstrates how an LLM coding agent can systematically find, exploit, and remediate web application vulnerabilities using [OWASP Juice Shop](https://github.com/juice-shop/juice-shop) as the intentionally-vulnerable target.
What this repo contains
| Path | Purpose |
|---|---|
.devcontainer/ |
Dev-container definition (Dockerfile + post-create hook). Installs the full security toolkit and clones Juice Shop. |
.agents/skills/blackbox-pentest/ |
Black-box / DAST skill: recon → spidering → automated scanning → manual OWASP Top 10 tests. |
.agents/skills/whitebox-code-review/ |
White-box / SAST skill: SAST (semgrep, njsscan) + SCA + secret scanning + manual class-by-class review. |
.agents/skills/security-assessment-orchestrator/ |
Orchestrator: runs both skills, correlates/de-dupes findings, writes a consolidated report. |
PLAN_build-demo-environment.md |
Architecture/design decisions and implementation phases. |
SECURITY_CHECK_*.md |
Generated assessment reports (example outputs). |
`juice-shop/`, `assess-out/`, and `bak/` are **git-ignored** — they live only in the local working tree and are never committed.
⚖️ Authorized-use notice
**All offensive-security tooling in this project is for authorized testing only.**
Only run these skills and tools against:
- The local Juice Shop instance started by
post-create.sh(default and intended target). - Systems you own or are explicitly authorized in writing (signed contract, scope agreement, or applicable law) to test.
You are solely responsible for ensuring your use is legal. Unauthorized scanning or exploitation is illegal in most jurisdictions and can cause real damage.
Quickstart
1. Open in the dev-container
Open this repository in VS Code, then **Reopen in Container** (Dev Containers extension). `post-create.sh` automatically:
- Clones `juice-sho
Related Skills
Web Security Audit Skill
Skill de auditoria de seguridad web completa (pentest black-box). 70+ tests automatizados, OWASP Top 10, CVEs,
Pentest Swarm AI
Autonomous penetration testing using a swarm of AI agents. Orchestrates recon, classification, exploitation, a
Talon
Penetration Testing MCP for Claude Code. AI-assisted security testing with automated recon, service enumeratio
Claude Redteam
Agentic pentest/red-team framework orchestrated by Claude Code — automated recon, hunt, validate, and report w
Coverage Checklist
Instantiate the internal-network pentest coverage checklist for an engagement. Copies the master template (met
Audit Infra
Infrastructure-first security audit — secrets, supply chain, CI/CD, LLM/skill security, OWASP, STRIDE. Complem
Related Agents
Go Security Specialist
Expert Go security auditor specializing in vulnerability detection, secure coding practices, and security comp
Worker Validator
Black-box runtime validator (#256). Given a PR + its validation-contract.md, RE-RUNS each assertion's command
Smoke Test Writer
Escribe smoke tests black-box contra el entorno dev desplegado. Asume que el proyecto SmokeTests ya existe.