The cve.now MCP server is an implementation of the Model Context Protocol that connects your AI agent and assistants like Claude, Cursor, etc directly to cve.now. It provides structured and secure access to vulnerability feeds, vulnerability records, and package-version checks, so your agent can check dependencies, look up CVE, GHSA, and OSV entries, review the newest feed items, and find vulnerabilities affecting a specific package version on your behalf.
- Dependency vulnerability checks: Have your agent check a lockfile or package list against cve.now's indexed catalog and identify matching vulnerabilities for exact package versions.
- Vulnerability lookups: Direct your agent to retrieve a vulnerability entry by its CVE, GHSA, or OSV reference.
- Latest vulnerability feeds: Let the agent review new items from the critical, all, or KEV feeds to help track recently published risks.
- Technology and ecosystem monitoring: Instruct your agent to list feed items for a specific technology or package ecosystem.
- Responsible result interpretation: Have your agent explain that no catalog matches do not guarantee safety and distinguish model-written summaries from upstream vulnerability data.