Citation Link Validator

• "generate report with references"
• "cite sources and ensure links are valid"
• "no broken links"
• "validate all cited URLs"

Mode B: Post-validation Mode

When to Use:

• User uploads existing Markdown documents with footnotes
• Need to check link status in existing articles
• Regular document maintenance

Trigger Keywords:

• "check links in this document"
• "validate footnotes in my uploaded file"
• "which links are broken"

Mode A: Real-time Validation Workflow

Core Principles

Core Philosophy: Validate first, then add — never include unvalidated URLs in the final article

Workflow:

1. Use web_search to find relevant data
2. Extract candidate URLs from search results
3. Before adding footnotes, validate each URL for accessibility
4. Only add verified URLs to footnotes
5. If URL is broken, immediately search for alternative sources
6. Ensure all footnotes in final article are valid links

Step-by-Step Guide

Step 1: Search and Obtain Candidate URLs

# Use web_search to find relevant topics
search_results = web_search("AI ethics research site:edu OR site:org")

# Extract URLs from results
candidate_urls = [result['url'] for result in search_results]

Step 2: Real-time Validation of Each URL

Use the check_single_url.py script to validate individual URLs:

python /mnt/skills/user/citation-link-validator/scripts/check_single_url.py "<URL>"

Return Values:

• Exit code 0: URL is valid (HTTP status code < 400)
• Exit code 1: URL is broken or inaccessible

Python Call Example:

import subprocess

def is_url_valid(url: str, timeout: int = 10) -> bool:
    """Validate if a single URL is valid"""
    result = subprocess.run(
        ['python', '/mnt/skills/user/citation-link-validator/scripts/check_single_url.py',
         url, '--timeout', str(timeout), '--quiet'],
        capture_output=True
    )
    return result.returncode == 0

# Usage example
url = "https://www.nature.com/articles/s41586-023-12345-6"
if is_url_valid(url):
    print(f"URL is valid: {url}")