VulnClaw

VulnClaw is an autonomous infiltration testing tool based on LLM. MCP automates information gathering, vulnerability detection, abuse and reporting processes with tool chain and customizable skill sets. The system manages infiltration testing operations with evidence-based verification mechanisms and structured data processing.

VU

Netw0rkNoob · Security & Penetration Testing

STARS3.8K
FORKS535
COMMITS0
CONTRIBUTORS0
VERSIONv0.4.1
UPDATEDyesterday
PythonPenetration TestingLLM AgentMCPMIT

Common commercial products in this category: Burp Suite Professional, Nessus, Cobalt Strike

What can it do?
  • You can start automated infiltration testing processes by entering natural language commands.
  • You can detect vulnerabilities in web applications and create verifiable PoC scripts.
  • You can use autonomous agent cycles in CTF competitions or authorized safety tests.
User reviews
  • ↓ NegativeGitHub Issue

    [Feature]: reviewed cross-session experience — distill lessons from runs, gate them on human approval

    “## Problem Every run starts from zero. Whatever the agent learns during a session — that a target's login is rate-limited, that a particular parameter is reflected but not exploitable, that twenty tool calls down one path produced nothing — is discarded when the session ends.…”

My note

Alternative pages

More in: Security & Penetration Testing