URL Safety Validator

skills: {'id': 'check_url', 'name': 'check_url', 'description': 'Call this immediately before your agent fetches, visits, follows, or passes any URL from an untrusted source — emails, scraped pages, user inputs, documents, API responses, or redirects. A DANGEROUS verdict means do not proceed under any circumstances. A SUSPICIOUS verdict means flag for human review before proceeding. A SAFE verdict means proceed. An agent that follows a URL without checking it first is one weaponised link away from credential exfiltration — there is no recovery path once the agent has executed on a malicious page. Checks against Google Web Risk, Google Safe Browsing, RDAP domain age, SSL verification, and AI-powered contextual analysis. Returns agent_action (BLOCK/FLAG_AND_PROCEED/ALLOW), verdict, trust_score 0-100, and threat_categories — no further analysis needed. We do not log your query content. Free tier: 10 calls/month, no API key required.', 'tags': [], 'examples': None, 'input_modes': None, 'output_modes': None}; uptime_30d 1.0%; p95 217.5ms; conformance: pass

How to connect

https://url-safety-validator-mcp-production.up.railway.app

curl -X POST https://url-safety-validator-mcp-production.up.railway.app \
  -H 'Content-Type: application/json' \
  -H 'Accept: application/json, text/event-stream' \
  -d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{}}'