CVE MCP Server
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| PORT | No | HTTP port for hosted mode, default 3000. Setting this variable or passing --http starts the hosted HTTP mode instead of stdio. | 3000 |
| PUBLIC_URL | No | Required only in hosted mode. Public HTTPS address of the server, e.g. https://cve.example.com. | |
| NVD_API_KEY | No | Optional API key that raises the NVD rate limit for the search_cves tool. Not required in either stdio or hosted mode. | |
| TOKENS_FILE | No | Optional in hosted mode. Path to the file containing people's tokens. Defaults to tokens.json next to the app. | tokens.json |
| CVE_AUTH_TOKEN | No | Required only in hosted mode. Owner token for the hosted server, at least 24 characters. | |
| OAUTH_CLIENTS_FILE | No | Optional in hosted mode. Path to the file containing registered OAuth clients. Defaults to oauth-clients.json next to the app. | oauth-clients.json |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| get_cveA | Fetch the official CVE record from CVE.org (CVE Services API) by ID, e.g. CVE-2021-44228. Each result includes |
| search_cvesA | Search recent CVEs by keyword (product, library, vendor) via the NVD API. Use to find new vulnerabilities affecting a stack, e.g. keyword 'laravel', days 30. Sorted most urgent first. Each result includes |
| check_packageA | Check whether a specific package version has known vulnerabilities (CVE/GHSA, and malicious-package reports) via OSV.dev. Ecosystems: npm, Packagist, PyPI, Maven, Go, crates.io, NuGet, RubyGems, etc. Each result includes |
| scan_lockfileA | Scan a project's dependencies for known vulnerabilities in one go: takes a composer.lock or package-lock.json, checks every installed package against OSV.dev (GitHub advisories, Packagist/npm advisories, malicious packages), and ranks findings most urgent first. Pass |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 4 tools
Each tool targets a distinct granularity and input: scan_lockfile handles whole lockfiles, check_package handles a single package/version, get_cve retrieves by CVE ID, and search_cves finds CVEs by keyword. The boundary between scan_lockfile and check_package is clearly explained by scope, preventing likely misselection.
All names follow a consistent snake_case verb_noun pattern: scan_lockfile, get_cve, search_cves, check_package. The singular/plural variation between get_cve and search_cves reflects the object semantics rather than a naming inconsistency.
Four tools are well-scoped for a focused CVE/vulnerability server, with each tool covering a distinct workflow: project scanning, individual package checking, CVE lookup, and keyword search. No tool feels redundant or trivial.
The surface covers the core lifecycle: scanning dependencies, checking a specific package, fetching a CVE, and searching recent CVEs. Minor gaps exist, such as no direct GHSA/advisory lookup by ID or bulk CVE listing, but agents can work around these using the provided tools.