ai-scanner

Discover and classify public Ollama and ComfyUI instances via Shodan.

1|Updated Feb 1, 2026
One-click install
npx skills add https://github.com/safzanpirani/pi-config --skill ai-scanner
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ai-scanner
Source: https://github.com/safzanpirani/pi-config/tree/main/skills/ai-scanner
Command: npx skills add https://github.com/safzanpirani/pi-config --skill ai-scanner

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires shodan-api, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill provides passive discovery and probing of publicly exposed AI hosting instances, allowing users to scan and verify open AI-hosting hosts, ComfyUI instances, or Ollama servers.

Core Features & Use Cases

  • Shodan-based Discovery: Utilizes Shodan to identify and list publicly exposed AI hosting instances.
  • Classification & Caching: Classifies discovered instances and caches their IPs for future reference.
  • Optional GPU Details: Can optionally enrich instance details with GPU information for better instance selection.
  • Use Case: When looking for AI hosting services with specific GPU capabilities, the Skill can scan and report back on instances with matching GPUs.

Quick Start

Run the ai-scanner skill with the provided SHODAN_COOKIE to discover Ollama instances with 'deepseek-v4-pro' models.

Frequently Asked Questions about ai-scanner

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I discover publicly exposed Ollama instances using Shodan?

The ai-scanner performs passive Shodan-based probing to discover publicly exposed Ollama instances. It requires Shodan API access to query and identify these AI hosting servers across cloud environments.

What is Shodan-based AI service inventory and how does it work?

Shodan-based AI service inventory is the passive discovery of publicly exposed AI hosting instances. The ai-scanner queries Shodan, classifies discovered instances, and caches their IPs for future reference during security audits.

Do I need Shodan API access and cookies to scan for ComfyUI instances?

Yes, you need Shodan API access for discovery and Shodan cookies for web scraping to scan for ComfyUI instances. These credentials allow the ai-scanner to perform passive probing and retrieve optional GPU information.

Can I find AI hosting instances with specific GPU capabilities?

Yes, you can find AI hosting instances with specific GPU capabilities. The ai-scanner optionally enriches discovered Ollama and ComfyUI instances with GPU information to support better instance selection and resource discovery.

What's the best way to perform a passive security audit of exposed AI servers?

The best way to passively audit exposed AI servers is using the ai-scanner, which queries Shodan to identify, classify, and cache IPs of publicly exposed Ollama and ComfyUI instances without active exploitation.

What are the limitations of using Shodan probing for AI service discovery?

A limitation of Shodan probing for AI service discovery is its passive nature, relying on Shodan API access and cookies for web scraping. It cannot detect instances not indexed by Shodan or bypass authentication barriers.