vuln-check
Allows scanning container images (e.g. python:3.12-slim) pulled directly from a registry, inspecting their OS and application packages for vulnerabilities and returning prioritized findings with fix/upgrade targets.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@vuln-checkscan this repo for vulnerabilities and tell me what to upgrade"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
vuln-check
Vulnerability truth for AI agents. Runs grype locally, merges an OSV freshness overlay, and returns a prioritised answer with concrete upgrade targets instead of a 40 KB JSON dump.
Any language with a lockfile (npm, PyPI, Go, Maven, Cargo, RubyGems, Composer, NuGet, Hex, Pub, Swift, R, Conan…), OS packages, container images, SBOMs.
One score per finding: CVSS + EPSS + CISA KEV + fix availability.
Remediation plan per package: smallest version that closes every finding,
breakingflag.CLEANis ~20 tokens. Nothing runs on a server; nothing leaves your machine except the optional OSV query.
Install
# grype is required
brew install grype # or: curl -sSfL https://get.anchore.io/grype | sh -s -- -b /usr/local/bin
go install github.com/ciloops/vuln-check/cmd/vuln-check@latest
# or
brew install ciloops/tap/vuln-check # macOS cask
# or
npx @ciloops/vuln-check scan .Binaries for Linux/macOS/Windows on the Releases page. Docker: ghcr.io/ciloops/vuln-check (bundles grype).
Related MCP server: cve-cache-mcp
CLI
vuln-check scan . # agent format, OSV overlay on
vuln-check scan python:3.12-slim # image, pulled from the registry
vuln-check scan sbom:bom.cdx.json
vuln-check scan --format sarif --sarif-out out.sarif .
vuln-check scan --fail-above 69 --no-osv . # CI gate, exit 2 above threshold
vuln-check vuln CVE-2024-3094
vuln-check db updateOutput (--format agent):
{"status":"ACTION_REQUIRED","target":"dir:.","packages":412,"findings":17,"shown":5,"sources":["grype","osv"],
"actions":[{"package":"pkg:npm/minimist","current":"1.2.5","target":"1.2.6","breaking":false,"fixes":["CVE-2021-44906"],"score":59}],
"top":[{"id":"CVE-2021-44906","package":"minimist@1.2.5","severity":"CRITICAL","score":59,"epss":0.015,"fixed_in":["1.2.6"]}],
"unfixable":3}status: CLEAN | REVIEW (all findings < 40) | ACTION_REQUIRED. fresh: true marks advisories that came from OSV before grype-db ingested them. unfixable counts findings with no fix on the current release line.
MCP
Tools: scan_dir, scan_image, scan_sbom, vuln_lookup, db_update.
Claude Code:
claude mcp add vuln-check -- vuln-check mcp
# or the plugin (adds a skill that tells Claude when to scan):
claude plugin marketplace add ciloops/vuln-check && claude plugin install vuln-checkCursor / Windsurf / VS Code (.cursor/mcp.json, ~/.codeium/windsurf/mcp_config.json, .vscode/mcp.json):
{ "mcpServers": { "vuln-check": { "command": "vuln-check", "args": ["mcp"] } } }Codex (~/.codex/config.toml):
[mcp_servers.vuln-check]
command = "vuln-check"
args = ["mcp"]Add --no-osv to the args to keep the dependency list off the network.
GitHub Action
- uses: ciloops/vuln-check@v1
with:
target: . # or an image ref
fail-above: 69
sarif: vuln.sarif
- uses: github/codeql-action/upload-sarif@v3
if: always()
with: { sarif_file: vuln.sarif }Score
severity = CVSS base × 6 (label fallback) · exploit = EPSS × 30 · KEV: +10, floor 85 · no fix: × 0.85. Clamped to 0–100. See internal/score.
Freshness
grype-db builds daily and grype refreshes it at most every 2 h. The OSV overlay queries api.osv.dev with the purls grype inventoried (language ecosystems only) and adds anything grype-db has not seen yet, usually minutes after a GHSA is published. Disable with --no-osv.
Limits
No lockfile → transitive dependencies are invisible. Same as every scanner.
Remediation plans assume a linear release line per package; direct vs transitive is not distinguished yet.
Version comparison is a generic tokenizer, not per-ecosystem parsers. Good enough for "smallest fix above current"; edge cases exist (PEP 440 post-releases, exotic Maven qualifiers).
Development
go test ./...
go build ./cmd/vuln-checkTests run against a fake grype (cmd/vuln-check/testdata/fake-grype.sh) and fixtures in internal/grype/testdata. CI runs a real grype e2e.
Release: tag vX.Y.Z → goreleaser publishes binaries, checksums, ghcr image, Homebrew cask for macOS (needs HOMEBREW_TAP_GITHUB_TOKEN and the ciloops/homebrew-tap repo) and the npm package via trusted publishing (OIDC, no token: configure the GitHub Actions trusted publisher for @ciloops/vuln-check on npmjs.com, workflow release.yml).
License
Apache-2.0. grype and syft are Apache-2.0 by Anchore. Vulnerability data: NVD (public domain), GHSA (CC-BY 4.0), OSV (per source), CISA KEV, FIRST EPSS.
This server cannot be deployed
Maintenance
Related MCP Connectors
CVE lookups (NVD) and dependency-manifest audits (OSV) for AI agents. No API keys.
CVE lookups (NVD) and dependency-manifest audits (OSV) for AI agents. No API keys.
Look up CVEs, CISA KEV, EPSS and open-source package vulnerabilities, including malicious packages.
Generate SBOMs, scan vulnerabilities, and analyze dependencies from local projects or Git repos.
Related MCP Servers
- AlicenseAqualityBmaintenanceVulnerability intelligence for AI agents that enables CVE lookup, package vulnerability checks, and dependency auditing without API keys.5MIT
- AlicenseAqualityDmaintenanceProvides cached, structured snapshots of recent CVEs and GitHub Security Advisories, enabling AI agents to lookup CVEs, find vulnerabilities for a package by ecosystem, list recent critical CVEs, and get severity summaries.422 npmApache 2.0
- AlicenseNot gradedqualityCmaintenanceProvides security checks for AI agents, including secret scanning, CVE lookup, and dependency vulnerability scanning, all running locally except for CVE lookups.MIT
- AlicenseAqualityBmaintenanceEnables AI agents to search, correlate, and monitor real-time vulnerability intelligence from NVD, CISA KEV, EPSS, and MITRE ATT&CK, including CVE details, critical CVE tracking, known exploited vulnerabilities, and exploitation probability scores.52MIT