pravidhi-ddgs-internet-search

Perform web searches using DuckDuckGo. Features soft-failure mode, generic environment support, and context-safe output.

v1.0.0 MIT

Internet Search Skill (DDGS)

​> Developed and maintained by Pravidhi

This skill performs internet searches using the DuckDuckGo Search library. It is designed to be low-friction and context-safe for AI agents.

Prerequisite: Installation

This skill requires the duckduckgo-search library. Install it in your agent’s active Python environment.

bash
pip install duckduckgo-search

Usage

Standard Search: Execute the script using your active Python interpreter.

bash
# Syntax: python scripts/search.py -q "QUERY" -m MAX_RESULTS

python scripts/search.py -q "AI funding 2026" -m 3

Safe Mode (Use if getting timeouts): The script includes a --safe flag that adds extra delay (3s) and disables aggressive backends.

bash
python scripts/search.py -q "latest linux kernel features" -m 5 --safe

Known Limitations & Troubleshooting

​> Warning: This skill relies on scraping DuckDuckGo. It is not an official API.

  1. Rate Limiting / IP Blocking:

    • Symptom: 429 Too Many Requests or Ratelimit errors.
    • Cause: Cloud/Datacenter IPs (AWS, GCP) are often flagged by DuckDuckGo.
    • Fix: Wait 60 seconds. Try the --safe flag. If persistent, use a residential proxy or switch to an API-based skill.

  2. Soft Failures:

    • The script is designed to “soft fail”. Instead of crashing with a non-zero exit code, it will return a JSON object with status: error. Agents should parse the JSON output to handle errors gracefully.

References