Skip to content

Orizon-eu/claude-code-pentest

claude-code-pentest

6 Claude Code skills that automate the entire pentest lifecycle.
From recon to exploit chains to bug bounty reports — just give it a domain.

Python 3.8+ License: MIT Claude Code 6 Skills 43 Scripts Zero Dependencies

Built by orizon.one


WARNING: Authorized security testing only. Unauthorized access to computer systems is illegal. Read the full disclaimer before use.


The Pipeline

DOMAIN INPUT
    │
    ▼
┌─────────────────────┐
│  recon-dominator     │  Subdomain enum, port scan, OSINT, dorking, Wayback
└────────┬────────────┘
         │
    ┌────┼──────────────────────┐
    │    │                      │
    ▼    ▼                      ▼
┌────────────┐ ┌──────────┐ ┌──────────────┐
│ webapp-    │ │ api-     │ │ cloud-pivot- │
│ exploit-   │ │ breaker  │ │ finder       │
│ hunter     │ │          │ │              │
└─────┬──────┘ └────┬─────┘ └──────┬───────┘
      │              │              │
      └──────┬───────┘──────────────┘
             ▼
┌─────────────────────────┐
│  attack-path-architect  │  MITRE ATT&CK trees, kill chains
└────────┬────────────────┘
         ▼
┌─────────────────────────┐
│  vuln-chain-composer    │  Chain exploits → bug bounty report
└─────────────────────────┘

What Each Skill Does

# Skill What It Does Scripts
1 recon-dominator Subdomain enumeration, port scanning, tech fingerprinting, OSINT, Google dorking, Wayback analysis 8
2 attack-path-architect Asset classification, trust mapping, MITRE ATT&CK attack trees, kill chain generation 3
3 webapp-exploit-hunter SQLi, XSS, SSRF, IDOR, SSTI, auth bypass, file upload, race conditions + PoC generation 11
4 api-breaker API discovery, schema reconstruction, BOLA/BFLA, mass assignment, JWT attacks, GraphQL abuse 8
5 cloud-pivot-finder Cloud provider detection, S3/GCS/Azure buckets, subdomain takeover, serverless, CI/CD exposure 7
6 vuln-chain-composer Cross-domain vuln correlation, exploit chain composition, CVSS recalculation, bug bounty reports 6

43 Python scripts total. Zero pip dependencies. Pure standard library.

Quick Start

# Clone
git clone https://github.com/Orizon-eu/claude-code-pentest.git
cd claude-code-pentest

# Install all skills (personal - available in all your Claude Code projects)
for skill in recon-dominator attack-path-architect webapp-exploit-hunter api-breaker cloud-pivot-finder vuln-chain-composer; do
    cp -r "$skill" ~/.claude/skills/
done

# Verify
# Open Claude Code and ask: "What skills are available?"

Then just talk to Claude:

Run full reconnaissance on example.com
Find vulnerabilities on app.example.com
Chain all findings and generate a bug bounty report

Requirements

  • Python 3.8+ (standard library only)
  • Claude Code with skills support

Optional (for faster scanning)

go install -v github.com/projectdiscovery/subfinder/v2/cmd/subfinder@latest
go install -v github.com/projectdiscovery/httpx/cmd/httpx@latest
go install github.com/d3mondev/puredns/v2@latest
go install -v github.com/projectdiscovery/naabu/v2/cmd/naabu@latest

System tools (usually pre-installed): dig, whois, nmap

Installation

Method 1: Personal skills (all projects)

git clone https://github.com/Orizon-eu/claude-code-pentest.git
cd claude-code-pentest

for skill in recon-dominator attack-path-architect webapp-exploit-hunter api-breaker cloud-pivot-finder vuln-chain-composer; do
    cp -r "$skill" ~/.claude/skills/
done

Method 2: Project-level skills (single project)

git clone https://github.com/Orizon-eu/claude-code-pentest.git
cd your-project

mkdir -p .claude/skills
for skill in recon-dominator attack-path-architect webapp-exploit-hunter api-breaker cloud-pivot-finder vuln-chain-composer; do
    cp -r "/path/to/claude-code-pentest/$skill" .claude/skills/
done

Method 3: Single skill

cp -r recon-dominator ~/.claude/skills/

Usage

Natural language (Claude auto-selects the right skill)

Run full recon on target.com
Scan all web apps for vulnerabilities
Test the API at api.target.com
Check for S3 buckets and subdomain takeover
Analyze attack paths from the recon data
Chain all findings and generate a report

Direct invocation

/recon-dominator
/attack-path-architect
/webapp-exploit-hunter
/api-breaker
/cloud-pivot-finder
/vuln-chain-composer

Full Pentest Workflow

1. "Run full recon on target.com"
   → recon-dominator maps the entire attack surface

2. "Scan all web apps for vulnerabilities"
   → webapp-exploit-hunter tests for SQLi, XSS, SSRF, SSTI, IDOR...

3. "Test all discovered APIs"
   → api-breaker finds BOLA, BFLA, JWT issues, mass assignment

4. "Check cloud infrastructure and buckets"
   → cloud-pivot-finder maps cloud pivot paths

5. "Analyze attack paths from all data"
   → attack-path-architect generates MITRE ATT&CK kill chains

6. "Chain all findings and generate a report"
   → vuln-chain-composer produces bug bounty ready reports

Each skill outputs structured JSON that feeds into the next skill in the pipeline.

Detailed Script Reference

recon-dominator — 8 scripts
Script Function
passive_recon.py crt.sh, HackerTarget, RapidDNS, Wayback, subfinder integration
active_recon.py DNS brute-force, zone transfer, permutation scanning
port_scanner.py TCP connect scan, HTTP probing, WAF detection
tech_fingerprint.py 50+ technology signatures, security header analysis
osint_correlator.py WHOIS, ASN, reverse IP, GitHub dorks, email providers
google_dorker.py 88 dorks across 9 categories
wayback_analyzer.py URL categorization, JS analysis, removed content detection
generate_report.py Consolidated MD + JSON report
attack-path-architect — 3 scripts
Script Function
classify_assets.py Asset type classification, exposure level, risk scoring
map_trust.py Cookie scope, DNS relationships, TLS sharing, SSO detection
generate_attack_tree.py Attack patterns, MITRE TTP mapping, kill chain generation
webapp-exploit-hunter — 11 scripts
Script Function
crawler.py HTML parsing, form extraction, JS endpoint discovery, path fuzzing
sqli_tester.py Error-based, boolean-based, time-based, UNION-based SQLi
xss_tester.py Context-aware XSS (HTML, attribute, JS, URL), WAF bypass
ssrf_tester.py Localhost, cloud metadata, protocol smuggling, IP bypass
ssti_tester.py Jinja2, Twig, Freemarker, ERB, Smarty, Pebble, Velocity
idor_tester.py Horizontal/vertical access, method variation, ID encoding
auth_tester.py Default creds, JWT attacks, session analysis, CORS, headers
upload_tester.py Double extension, null byte, magic bytes, Content-Type bypass
race_tester.py Double spend, coupon reuse, parallel request racing
generate_poc.py curl commands, Python scripts, reproduction steps
vuln_report.py Bug bounty reports with CVSS scoring
api-breaker — 8 scripts
Script Function
api_discovery.py 70+ common API paths, GraphQL introspection, OpenAPI parsing
schema_builder.py API schema reconstruction, OpenAPI 3.0 output
auth_analyzer.py JWT decode/attack, none algorithm, key confusion, weak secrets
authz_tester.py BOLA/BFLA testing with dual tokens
mass_assignment.py 30+ injection fields, nested object testing
rate_limiter.py 120+ rapid requests, 12 IP spoofing headers, GraphQL batching
logic_tester.py Price manipulation, quantity overflow, privilege escalation
api_report.py OWASP API Top 10 mapping, per-finding reports
cloud-pivot-finder — 7 scripts
Script Function
cloud_detector.py CNAME patterns, header analysis, IP whois for cloud providers
bucket_enum.py S3/GCS/Azure bucket name generation and permission testing
takeover_scanner.py 18 service fingerprints for dangling CNAME detection
serverless_finder.py Lambda, API Gateway, Cloud Functions, Cloud Run discovery
cicd_finder.py Jenkins, GitLab, GitHub Actions, Terraform, Docker/K8s configs
metadata_paths.py SSRF vectors, metadata-to-credential pivot chain mapping
cloud_report.py Consolidated cloud security report with remediation
vuln-chain-composer — 6 scripts
Script Function
import_findings.py Normalizes findings from all tools to common format
correlate.py 10 chain templates matching vulnerability combinations
build_chains.py Detailed chain construction with per-phase steps
calculate_impact.py CVSS v3.1 recalculation, amplification rules
generate_chain_poc.py Python scripts, curl sequences per chain
generate_report.py HackerOne, Bugcrowd, and generic pentest report formats

Architecture

claude-code-pentest/
├── README.md
├── LICENSE
├── DISCLAIMER.md
├── SECURITY.md
├── CONTRIBUTING.md
├── CODE_OF_CONDUCT.md
├── .gitignore
├── recon-dominator/
│   ├── SKILL.md
│   ├── scripts/           (8 scripts)
│   └── references/        (wordlists, tool setup)
├── attack-path-architect/
│   ├── SKILL.md
│   ├── scripts/           (3 scripts)
│   └── references/        (MITRE ATT&CK mapping)
├── webapp-exploit-hunter/
│   ├── SKILL.md
│   └── scripts/           (11 scripts)
├── api-breaker/
│   ├── SKILL.md
│   └── scripts/           (8 scripts)
├── cloud-pivot-finder/
│   ├── SKILL.md
│   └── scripts/           (7 scripts)
└── vuln-chain-composer/
    ├── SKILL.md
    └── scripts/           (6 scripts)

Troubleshooting

Problem Solution
Skill not triggering automatically Invoke directly with /skill-name
Scripts fail with "command not found" Install optional Go tools (see Requirements)
Permission denied on scripts Run chmod +x scripts/*.py inside the skill folder
Too many skills loaded Install only the skills you need
Rate limiting (429 errors) Scripts auto-adjust rate; use --delay flag

Legal Disclaimer

This software is provided for authorized security testing and educational purposes only.

By using this software, you agree that:

  1. You have explicit, written authorization to test any target system
  2. You accept full responsibility for your actions and any consequences
  3. The authors are not liable for any misuse, damage, or legal consequences
  4. You will comply with all applicable laws in your jurisdiction

Unauthorized access to computer systems is a criminal offense under the Computer Fraud and Abuse Act (CFAA), Computer Misuse Act, Budapest Convention on Cybercrime, and equivalent laws worldwide.

Read the full DISCLAIMER.md before using these tools.

Contributing

See CONTRIBUTING.md for guidelines.

License

MIT License - See LICENSE for details.

Security

Found a vulnerability in this project? See SECURITY.md.


Built by orizon.one — Offensive Security & AI Integration

About

6 Claude Code skills that automate the entire pentest lifecycle. From recon to exploit chains to bug bounty reports — just give it a domain. 43 scripts, zero dependencies, pure Python.

Topics

Resources

License

Code of conduct

Contributing

Security policy

Stars

Watchers

Forks

Packages

 
 
 

Contributors

Languages