# Netw0rkNoob/VulnClaw

基于 AI Agent + MCP 工具链 + 渗透 Skill 编排， 配合大语言模型， 自然语言输入 → 自动完成「信息收集 → 漏洞发现 → 漏洞利用 → 报告生成」全流程。

Repository: https://github.com/Netw0rkNoob/VulnClaw
Canonical: https://ross.abutalabs.com/products/vulnclaw
Homepage: https://unclecheng-li.github.io/vulnclaw.com
Language: Python
License: MIT
License Family: permissive
Topics: ai, ai-agent, ai-tools, ctf, cybersecurity, openclaw, penetration-testing, penetration-testing-tools, security-tools, skill
Last push: 2026-08-23T10:43:52+00:00

## Health v2 (maintenance only)
Score: 80/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 99, release rhythm 96, longevity 9
- inputs: {"age_days": 137, "days_push": 10, "days_rel": 24, "gap_med": 9.5, "n_releases_24m": 7}
- flags: young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2997, forks 401 (observed 2026-08-28T04:07:35.101744+00:00)

## What it is
VulnClaw is an AI-driven penetration testing CLI tool that combines an LLM agent, MCP toolchain, and curated pentest skills to automate the full workflow from reconnaissance to vulnerability discovery, exploitation, and report generation from natural-language input. It supports 14 OpenAI-compatible LLM providers, evidence-level anti-hallucination checks, and CLI/REPL, TUI, Web UI, and Docker modes.

## Use cases
- run an authorized penetration test on a target website from a natural language prompt
- automate CTF flag hunting with AI
- generate structured pentest reports with PoC scripts
- perform reconnaissance, subdomain enumeration, and directory brute-forcing automatically
- orchestrate MCP tools like Burp and Chrome DevTools for web exploitation
- teach security students AI-assisted offensive testing workflows

## When to choose
- you want natural-language-driven automation of authorized pentest or CTF workflows
- you need an agent that grounds conclusions in real tool output to avoid hallucinated results
- you want flexible LLM backend choice including local Ollama
- you need both terminal and web interfaces plus Docker deployment

## When to avoid
- you need a traditional deterministic vulnerability scanner with stable signatures
- your targets are not explicitly authorized for testing
- you cannot expose an LLM API key or local model to your testing environment
- you require compliance-audited enterprise scanning tooling

## Facets
- artifact type: cli-tool
- maturity: active
- function: agent-framework, mcp, penetration-testing, llm-inference, cli, chatbot
- domain: penetration-testing, security, artificial-intelligence, developer-tools
- platform: cli, python, cross-platform, self-hosted
- tags: ai-pentest, mcp-toolchain, ctf, red-team, vulnerability-scanning, report-generation, anti-hallucination, llm-agent, offensive-security, ai-agents, command-line, docker

## Member repositories
- Netw0rkNoob/VulnClaw (main) score 80

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:35.101744+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T07:30:43.931699+00:00, confidence not recorded.
  - readme: https://github.com/Netw0rkNoob/VulnClaw (fetched 2026-08-28T04:07:35.101744+00:00, sha d94bb105becc)
  - homepage: https://unclecheng-li.github.io/vulnclaw.com (fetched 2026-08-29T09:45:56.953611+00:00, sha d661c6320365)
- Data as of 2026-08-30T08:39:29.467469+00:00.