# NPMScan > NPMScan is a threat intelligence platform for the npm ecosystem: package risk analysis, install-script inspection, and a searchable, category-filterable feed of known vulnerabilities and security advisories. NPMScan also runs a public, read-only **MCP (Model Context Protocol) server** so AI agents can look up npm package metadata and vulnerabilities directly instead of scraping the site or guessing from a package's README. No API key required. ## MCP Server (for AI agents) - [MCP endpoint](https://npmscan.com/api/mcp): Streamable HTTP, JSON-RPC 2.0. Add with `claude mcp add --transport http npmscan https://npmscan.com/api/mcp` or any MCP-compatible client (Claude Code, Claude Desktop, Cursor, ChatGPT Developer Mode). - [ChatGPT app](https://chatgpt.com/plugins/plugin_asdk_app_6a6a699e6f3481918d5e6034432894f2?q=npm): npmscan is an approved app in the ChatGPT Apps directory — one-click install instead of adding the connector manually. - [Claude directory listing](https://claude.ai/customize/skills/id/aa8fcafe-0bcd-41b6-9e07-6c7ebbb71153%40anthropic-plugin-directory): npmscan is officially listed in Anthropic's Claude plugin directory — add it from claude.ai in one click. - [Claude Code plugin](https://github.com/salemalem/npmscan-mcp-plugin): wraps this MCP server with five bundled skills: `dependency-audit`, `package-trust-check`, `new-dependency-evaluation`, `incident-response`, and `ci-pr-gate`. - ClawHub skills: the same five skills, published individually — [npmscan-dependency-audit](https://clawhub.ai/salemalem/skills/npmscan-dependency-audit), [npmscan-package-trust-check](https://clawhub.ai/salemalem/skills/npmscan-package-trust-check), [npmscan-new-dependency-evaluation](https://clawhub.ai/salemalem/skills/npmscan-new-dependency-evaluation), [npmscan-incident-response](https://clawhub.ai/salemalem/skills/npmscan-incident-response), [npmscan-ci-pr-gate](https://clawhub.ai/salemalem/skills/npmscan-ci-pr-gate). - [MCP setup guide](https://npmscan.com/docs/mcp): Connection instructions and config snippets. - [MCP tools reference](https://npmscan.com/mcp): Input schema plus real request/response examples for all 23 tools (get_package, get_package_version, get_maintainer_profile, search_packages, query_vulnerabilities, batch_query_vulnerabilities, get_latest_advisories, get_cve, analyze_install_script, analyze_transitive_dependencies, check_package_provenance, check_maintainer_changes, check_maintainer_blast_radius, check_license_compliance, diff_dependencies, prioritize_remediation, simulate_dependency_upgrade, suggest_alternative, compare_packages, audit_github_repository, get_remediation_playbook, generate_sbom, enrich_npm_audit). ## GitHub Action (for CI on pull requests) - [Setup guide](https://npmscan.com/docs/github-action): Step-by-step setup, every input/output, copy-paste rules for AI coding assistants, and a real request/response. - [GitHub Marketplace listing](https://github.com/marketplace/actions/npmscan-dependency-check): `npmscan/npmscan-action@v1` — scans package.json / package-lock.json / yarn.lock / pnpm-lock.yaml changes in a PR, posts a PR comment, and fails the check on vulnerable packages, new install scripts, or changed package sources. No API key. - [Source code](https://github.com/NPMscan/npmscan-action) - Minimal workflow: add `.github/workflows/npmscan.yml` with `on: pull_request`, `permissions: { contents: read, pull-requests: write }`, and steps `actions/checkout@v7` then `npmscan/npmscan-action@v1` (set `with: { file: pnpm-lock.yaml | yarn.lock | package.json }` if there is no package-lock.json). Optional inputs: `mode: warn` to report without blocking, `fail-on-severity` (low|moderate|high|critical|none), `fail-on-install-script`, `fail-on-source-change`, `fail-on-error`, and `files` for several lockfiles (monorepos). ## Docs - [Docs home](https://npmscan.com/docs) - [REST API](https://npmscan.com/api) ## Site - [Search packages](https://npmscan.com/search) - [Analyze a dependency tree](https://npmscan.com/analyze) - [Latest vulnerabilities](https://npmscan.com/latest-vulnerabilities): Filterable by severity, 20 vulnerability categories (XSS, SQL/NoSQL Injection, SSRF, Access Control, Code Injection, and more), and searchable by npm package name or an exact GHSA/CVE ID. - [Latest vulnerabilities RSS](https://npmscan.com/latest-vulnerabilities/rss.xml): Same feed with severity/category `` tags on every item for filtering in a feed reader.