IGS — Intelligence Gathering System v1.0.0

MCP server + CLI for intelligence gathering. 64 tools, 411 sources, 47 countries, TOON token-efficient output, Obscura headless browser.

Metric	Value
Tools	64 (56 core + 8 Obscura browser)
Intelligence Domains	20 (Discovery, News, Research, Web, Insights, Social, Twitter, Weather, Finance, Security, Patents, Government, Legal, Environment, Climate, Health, Politics, Browser, SOP, YouTube)
Sources	411 across 47 countries
Pools	14 (geopolitics, tech, India, defense, health, etc.)
Binary	Single igs binary (~26 MB musl static)
Output	TOON (default, ~40% fewer tokens) or JSON
API Keys	None required — all web search uses Obscura + DuckDuckGo

---

Installation

Option 1: Install Script (Recommended)

Detects your platform, downloads the latest release, and installs to ~/.local/bin:

curl -sSL https://raw.githubusercontent.com/ishan-parihar/igs-rust/master/scripts/install.sh | bash

Option 2: Manual Download

Download the tarball for your platform from the latest release, then:

tar -xzf igs-*.tar.gz
sudo mv igs /usr/local/bin/
igs --version

Option 3: Build from Source

# Prerequisites
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh
rustup target add x86_64-unknown-linux-musl

# Clone and build
git clone https://github.com/ishan-parihar/igs-rust.git
cd igs-rust
cargo build --release --target x86_64-unknown-linux-musl

# Install
sudo cp target/x86_64-unknown-linux-musl/release/igs /usr/local/bin/
igs --version

---

Quick Start

As MCP Server (for AI agents)

# Start the MCP server on stdio
igs mcp

Configure in Claude Desktop (~/.config/Claude/claude_desktop_config.json):

{
  "mcpServers": {
    "igs": {
      "command": "igs",
      "args": ["mcp"]
    }
  }
}

Configure in Cursor (.cursor/mcp.json):

{
  "mcpServers": {
    "igs": {
      "command": "igs",
      "args": ["mcp"]
    }
  }
}

Configure in OpenCode (~/.config/opencode/opencode.json):

{
  "mcp": {
    "igs": {
      "type": "local",
      "command": ["/usr/local/bin/igs", "mcp"],
      "enabled": true
    }
  }
}

As CLI

# System status
igs status

# Fetch news
igs news fetch --pools GLOBAL_TECH_CYBER --limit 10

# Search Reddit
igs reddit search --query "AI safety"

# Search academic papers
# Search web (zero API keys — uses DuckDuckGo via Obscura)
igs web search --query "rust async runtime"

# Scrape a URL to markdown
igs web scrape --url https://example.com

# Crawl a website (requires Obscura enabled)
igs web crawl --url https://example.com --max-depth 2

# Browser automation (requires Obscura enabled)
igs browser goto --url https://example.com
igs browser markdown
igs browser links
igs browser markdown
igs browser links

# List available pools, sources, parsers
igs pools list
igs sources list --pool GLOBAL_TECH_CYBER
igs sources countries
igs parsers

Output Format

All bulk data tools default to TOON (token-efficient). Use --format json for standard JSON:

igs --format json news fetch --pools GLOBAL_TECH_CYBER --limit 5
igs --format toon news fetch --pools GLOBAL_TECH_CYBER --limit 5

---

Configuration

Config Directory

IGS auto-creates ~/.config/igs-mcp/ on first run with default config files:

~/.config/igs-mcp/
├── settings.yml      # Main configuration
├── pools.yml         # 14 pool definitions
├── sources.yml       # 411 source definitions
├── countries.yml     # 47 count