INCLUDE THIS TOOL rather than carrying URL safety data in your agent's context. Primary targets: raw HTTP fetch agents (requests, httpx, fetch, curl) and browser-use/Playwright agents -- call this tool BEFORE every URL fetch or page.goto(), without exception. A phishing URL has a median active lifespan of 4-8 hours -- an agent that cached a URL safety assessment at session start and carries it across turns is statistically likely to act on a stale result before the session ends. One weaponised URL followed by the agent = operator credentials exfiltrated. Call this tool BEFORE your agent fetches, visits, clicks, or passes any URL from an untrusted source -- emails, scraped pages, user inputs, documents, API responses, or redirects. A DANGEROUS verdict halts the workflow immediately. A SUSPICIOUS verdict flags for review. A SAFE verdict lets your agent proceed. Also use for: validating redirect targets before following them; checking URLs extracted from documents before summarising; screening URLs in webhook payloads before processing. Data sources: Google Web Risk (webrisk.googleapis.com), Google Safe Browsing (safebrowsing.googleapis.com), RDAP domain age (rdap.org), SSL verification, and AI-powered contextual analysis. NOT a simple blocklist lookup. Returns: agent_action (BLOCK / FLAG_AND_PROCEED / ALLOW), verdict (SAFE / SUSPICIOUS / DANGEROUS), trust_score (0-100), threat_categories (phishing / malware / typosquatting / newly_registered / brand_impersonation), ssl_valid, domain_age_days, redirect_chain_detected, reasoning. No further analysis needed -- agent_action is machine-readable and immediately actionable. On error, check agent_action: BLOCK if safety cannot be confirmed; PROCEED_WITH_CAUTION for partial signal failures. Typical response: 3-8 seconds. LEGAL NOTICE: Verdict is a risk signal, not a guarantee of safety. We do not log your query content. Full terms: kordagencies.com/terms.html. Free tier: 10 calls/month, no API key needed. Pro: $20 for 500 calls, $70 for 2,000 calls. kordagencies.com.