What It Is

This server — the same VPS that hosts the site you're reading — doubles as a dedicated out-of-band (OOB) testing platform for bug bounty research. nginx is configured to log callbacks to specific endpoints, and a custom Python tool parses those logs to detect when a target application makes an outbound request to my server.

The server has been running continuously for five months.

The Offensive Use Case

Some of the most impactful web vulnerabilities are "blind" — the application processes your payload, but the result doesn't show up in the HTTP response you get back. You need a separate server listening for callbacks to confirm the vulnerability fired.

Blind XSS

In blind XSS testing, you inject a payload that includes a callback URL pointing to your OOB server. If the payload executes — maybe hours or days later, in an admin panel or internal dashboard that processes your input — the victim's browser makes a request to your server. The callback confirms execution, and the log captures context: the page URL, cookies (if not HttpOnly), and the user agent of whoever triggered the payload.

Blind SSRF

Server-Side Request Forgery testing works the same way. You submit a URL pointing to your OOB server in parameters that the application might fetch server-side (webhook URLs, import fields, avatar URLs, PDF generators). If the server makes the request, your callback log captures the source IP, the request path, and any headers the application forwards — confirming that the application can be directed to make outbound requests to arbitrary hosts.

The Defensive Parallel

The pattern behind this OOB server — collect network events into structured logs, then parse and analyze them — is the same pattern that drives defensive monitoring.

  • A SIEM ingests logs from firewalls, proxies, and endpoints, then correlates events to detect threats
  • An IDS watches network traffic and logs matches against known signatures
  • A SOC analyst reviews parsed log data to identify anomalous behavior

My OOB server does the same thing at a smaller scale: it collects inbound network events (HTTP callbacks), writes them to structured logs, and I parse those logs to extract the signal from the noise. The tooling is simpler, but the workflow is identical: collect, parse, correlate, act.

How It Works

  • nginx endpoint configuration: specific URL paths are configured as callback endpoints, each mapped to a different test case or target
  • Logging: nginx logs all requests to these endpoints with full request details — timestamp, source IP, request path, headers, and user agent
  • Python log parser: a custom tool that reads the nginx logs, filters for callback hits, extracts the relevant fields, and presents them in a readable format tied back to specific test payloads

Infrastructure

The OOB server runs on the same hardened VPS as this portfolio site:

  • Key-only SSH authentication
  • Firewall configured to expose only necessary ports
  • TLS for the portfolio site; callback endpoints accept both HTTP and HTTPS to catch payloads regardless of the target's outbound configuration
  • Regular OS and package patching

What This Demonstrates

  • Offensive tooling: purpose-built infrastructure for detecting blind vulnerabilities that can't be confirmed through normal HTTP responses
  • Log collection and parsing: the same collect-and-parse workflow used in security monitoring and SIEM operations
  • Infrastructure operation: maintaining a production server with continuous uptime, hardened access, and regular patching
  • Custom tooling: building specific tools to solve specific problems, rather than relying entirely on off-the-shelf solutions
All Projects