URL Safety Validator MCP
Have your agent label links SAFE, SUSPICIOUS, or DANGEROUS with trust scores before users or workflows follow them.
Overview
URL Safety Validator MCP is a MCP server for the ship phase that returns SAFE, SUSPICIOUS, or DANGEROUS verdicts with trust scores and threat intel for URLs.
What is this MCP server?
- Verdict taxonomy: SAFE, SUSPICIOUS, and DANGEROUS with trust score output
- Threat intelligence layered with AI trust scoring via Anthropic
- Optional Google Web Risk and Google Safe Browsing API keys with graceful degradation
- Streamable HTTP Railway remote and stdio npm package v1.2.6
- Fits agent browsing, webhook docs, and user-submitted URL review flows
- Server version 1.2.6
- Three verdict levels documented: SAFE, SUSPICIOUS, DANGEROUS
- Up to 3 API key integrations: Anthropic required, 2 optional Google keys
What problem does it solve?
Agents and apps that follow user-supplied or scraped URLs risk phishing, malware, and reputation damage without a fast verdict layer.
Who is it for?
Builders shipping agents with web fetch, bookmarking, support bots, or user-generated link features who need automated safety checks.
Skip if: Teams that only use fixed allowlisted domains and never process arbitrary external URLs.
What do I get? / Deliverables
After install, your agent can classify URLs with trust scores and threat signals before fetch, store, or redirect actions run.
- SAFE / SUSPICIOUS / DANGEROUS verdict per URL
- Numeric or structured trust score suitable for agent guardrails
- Threat-oriented context to block or warn before navigation
Recommended MCP Servers
Journey fit
How it compares
MCP URL reputation checker with optional Google intel, not a full WAF or browser isolation product.
Common Questions / FAQ
Who is URL Safety Validator MCP for?
It is for developers hardening agent and app workflows that encounter untrusted URLs and need structured safe-or-danger verdicts.
When should I use URL Safety Validator MCP?
Use it during ship security review and in production whenever links are submitted, scraped, or followed automatically.
How do I add URL Safety Validator MCP to my agent?
Register url-safety-validator-mcp with ANTHROPIC_API_KEY required; add Google Web Risk or Safe Browsing keys optionally; use stdio npm or the Railway HTTP remote.