{"name":"vulnti-mcp","version":"0.1.0","description":"Vulnti MCP server — vulnerability intelligence as agent tools. Point an MCP client at this URL (Streamable HTTP transport).","transport":"streamable-http","endpoint":"/mcp","tools":[{"name":"search_vulns","description":"Search vulnerabilities by software keyword (e.g. \"nginx\", \"openssl\", \"log4j\"). Matches CVE/GHSA ids, titles, summaries, and affected CPE product/vendor names. Optionally floor by minimum severity. Returns the most relevant recent matches."},{"name":"get_vuln","description":"Fetch a single vulnerability by identifier: CVE (CVE-YYYY-NNNN), GHSA, CNNVD, or the internal vulnti id. Returns full detail or found:false."},{"name":"recent_vulns","description":"List the most recently ingested vulnerabilities. Optionally floor by severity or restrict to KEV (CISA Known Exploited Vulnerabilities) entries."},{"name":"triage_cves","description":"PAID. Triage a vulnerability-scanner CVE list by real-world exploitability — the \"cleaning\" service for AWS Inspector / Nessus / Trivy / Qualys output. Give up to 200 CVE/GHSA ids; get each classified as exploited_in_wild (CISA KEV) / poc_available / unconfirmed / unknown, prioritized, with an actionable count so you patch what matters first. Requires an API key."}],"paid":{"triage":{"endpoint":"POST /triage","body":{"cve_ids":["CVE-YYYY-NNNN","…"]},"auth":"Authorization: Bearer <api-key>","description":"Exploitability triage / cleaning for scanner output (Inspector, Nessus, Trivy). Classifies each CVE as exploited_in_wild / poc_available / unconfirmed / unknown."}},"docs":"https://vulnti.com"}