Скачать или смотреть My AI Agent Exploited a CVE That Metasploit Couldn't — RedAmon Full Demo

I wanted to stress-test RedAmon — our autonomous security assessment agent — under the worst possible conditions. So I set up a deliberately vulnerable Node.js server running node-serialize 0.0.4 (CVE-2017-5941, CVSS 9.8) and gave the agent a single instruction:

"Find a CVE and exploit it."

Here's what made this brutally hard: 👇

❌ The recon database had zero CVEs — no vulnerability data at all
❌ Metasploit had no module for this CVE — search CVE-2017-5941 returned nothing
❌ The agent had to go from zero knowledge to full RCE completely on its own

🤖 Powered by Claude Opus 4.6, here's the autonomous attack chain RedAmon executed in ~20 steps:

[1] Queried the knowledge graph for known vulnerabilities — found nothing
[2] Queried for hosts, IPs, services — only a bare subdomain, no ports, no tech stack
[3] Launched a port scan (naabu) — discovered ports 22 (SSH) and 8080 (HTTP)
[4] Fingerprinted port 8080 via HTTP headers — identified Express/Node.js
[5] Fetched the full page source — spotted "Node.js 8.x" and "cookie-based serialization with Base64 encoding"
[6] Connected the dots: old Node.js + serialized cookies = classic deserialization attack surface
[7] Searched the web for CVE-2017-5941 exploit techniques and payloads
[8] Found multiple PoCs on Exploit-DB and PayloadsAllTheThings confirming the IIFE-based RCE vector
[9] Probed /profile endpoint — confirmed the server sets a Base64-encoded profile cookie
[10] Decoded the cookie: {"username":"guest","email":"[email protected]","role":"viewer"}
[11] Understood the full kill chain: cookie → Base64 decode → unserialize() → IIFE execution → RCE
[12] Requested phase transition from reconnaissance to exploitation — with a full risk assessment
[13] Searched Metasploit for a module — no results
[14] Pivoted to manual exploitation (no-module fallback)
[15] Crafted a malicious serialized payload with $$ND_FUNC$$ prefix and IIFE
[16] First attempt: async exec() — RCE fired server-side but output went to stdout, not HTTP response
[17] Self-corrected: switched to execSync() to capture output directly in the response
[18] Built the final payload embedding execSync('id') as the username field
[19] Base64-encoded the payload and sent it as the profile cookie
[20] RCE confirmed: the profile page rendered uid=0(root) gid=0(root) groups=0(root) as the username

🤯 Let that sink in.

The agent started with literally no vulnerability data, hit a dead end in Metasploit, researched the exploit autonomously on the web, crafted a custom deserialization payload, debugged its own approach when the first attempt didn't return output in the HTTP response, and achieved root-level RCE — all from a single sentence prompt.

🚫 No human guidance.
🚫 No hand-holding.
🚫 No pre-loaded exploit modules.

⚡ This is what autonomous offensive security looks like in 2026.

🔗 RedAmon is open source.
https://github.com/samugit83/redamon


⚠️ ETHICAL DISCLAIMER:
━━━━━━━━━━━━━━━━━━━━━
This tool is intended for AUTHORIZED security testing, educational purposes,
and research ONLY. Never scan or attack systems without explicit written
permission. Unauthorized access is ILLEGAL.

#AIRedTeam #PenetrationTesting #CyberSecurity #LangGraph #Metasploit
#Neo4j #ArtificialIntelligence #EthicalHacking #AgenticAI #MCP
#MachineLearning #InfoSec #BugBounty #OffensiveSecurity #Python
#TypeScript #Docker #Nuclei #OSINT #ReActPattern

🎓 About the Instructor:
I'm Samuele Giampieri, an AI engineer passionate about bridging cutting-edge research with practical applications. My expertise spans knowledge graphs, NLP, vector databases, and AI-driven retrieval systems, and I enjoy creating resources that empower innovation.

🔗 Connect with Me:
GitHub: https://github.com/samugit83
LinkedIn: /samuele-giampieri-b1b67597
Website: https://www.devergolabs.com

© 2026 | Educational Content for Authorized Security Testing Only