Hound MCP
The dependency bloodhound for AI coding agents. Hound is a free, open-source MCP server that gives AI coding agents a nose for supply chain security. It scans packages for vulnerabilities, checks licenses, inspects dependency trees, and detects typosquatting — with zero API keys, zero config, and zero cost.
README
Hound MCP
The dependency bloodhound for AI coding agents.
Hound is a free, open-source MCP server that gives AI coding agents a nose for supply chain security. It scans packages for vulnerabilities, checks licenses, inspects dependency trees, and detects typosquatting — with zero API keys, zero config, and zero cost.
<a href="https://glama.ai/mcp/servers/tiluckdave/hound-mcp"> <img width="380" height="200" src="https://glama.ai/mcp/servers/tiluckdave/hound-mcp/badge" alt="Hound MCP server" /> </a>
Why Hound?
Most security tools require accounts, API keys, or paid plans. Hound uses only two fully free, unauthenticated public APIs:
- deps.dev (Google Open Source Insights) — package metadata, dependency trees, licenses, OpenSSF Scorecard
- OSV (Google Open Source Vulnerabilities) — CVEs, GHSAs, fix versions
No sign-up. No config. Just install and go.
Quickstart
Claude Code
claude mcp add hound -- npx -y hound-mcp
Claude Desktop / Cursor / Windsurf
Add to your MCP config file:
{
"mcpServers": {
"hound": {
"command": "npx",
"args": ["-y", "hound-mcp"]
}
}
}
Config file locations:
- Claude Desktop (macOS):
~/Library/Application Support/Claude/claude_desktop_config.json - Cursor:
~/.cursor/mcp.json - Windsurf:
~/.codeium/windsurf/mcp_config.json
VS Code (Copilot)
{
"mcp": {
"servers": {
"hound": {
"type": "stdio",
"command": "npx",
"args": ["-y", "hound-mcp"]
}
}
}
}
Tools
Hound registers 12 tools in your MCP client.
hound_audit ⭐
Scan a whole project by passing your lockfile content. Parses package-lock.json, yarn.lock, pnpm-lock.yaml, requirements.txt, Cargo.lock, or go.sum and batch-queries OSV for vulnerabilities across all dependencies.
hound_audit(lockfile_name: "package-lock.json", lockfile_content: "<contents>")
hound_vulns
List all known vulnerabilities for a package version, grouped by severity with fix versions.
hound_vulns(name: "express", version: "4.18.2", ecosystem: "npm")
hound_inspect
Comprehensive package profile — licenses, vulnerabilities, OpenSSF Scorecard, GitHub stars, and dependency count in one call.
hound_inspect(name: "lodash", version: "4.17.21", ecosystem: "npm")
hound_score
Compute a 0–100 Hound Score combining vulnerability severity (40 pts), OpenSSF Scorecard (25 pts), release recency (20 pts), and license risk (15 pts). Returns a letter grade A–F with a full breakdown.
hound_score(name: "express", version: "4.18.2", ecosystem: "npm")
hound_upgrade
Find the minimum version upgrade that resolves all known vulnerabilities. Checks every published version and returns the nearest safe one.
hound_upgrade(name: "lodash", version: "4.17.20", ecosystem: "npm")
hound_compare
Side-by-side comparison of two packages across vulnerabilities, OpenSSF Scorecard, GitHub stars, release recency, and license. Returns a recommendation.
hound_compare(package_a: "express", package_b: "fastify", ecosystem: "npm")
hound_preinstall
Safety check before installing a package. Checks vulnerabilities, typosquatting risk, abandonment, and license. Returns a GO / CAUTION / NO-GO verdict.
hound_preinstall(name: "some-package", version: "1.0.0", ecosystem: "npm")
hound_tree
Full resolved dependency tree including all transitive dependencies, with depth and relation type.
hound_tree(name: "next", version: "14.2.0", ecosystem: "npm", maxDepth: 3)
hound_advisories
Full advisory details by ID — works with GHSA, CVE, and OSV IDs.
hound_advisories(id: "GHSA-rv95-896h-c2vc")
hound_advisories(id: "CVE-2024-29041")
hound_typosquat
Generates likely typo variants of a package name and checks which ones exist in the registry — surfaces potential typosquatting attacks.
hound_typosquat(name: "lodash", ecosystem: "npm")
hound_license_check
Scan a lockfile for license compliance. Resolves licenses for all dependencies and flags packages that violate the chosen policy.
hound_license_check(lockfile_name: "package-lock.json", lockfile_content: "<contents>", policy: "permissive")
Policies: permissive (MIT/Apache/BSD only), copyleft (allows GPL but not AGPL), none (report only).
hound_popular
Scan a list of popular (or user-specified) packages for known vulnerabilities. Great for a quick ecosystem health check.
hound_popular(ecosystem: "npm")
hound_popular(ecosystem: "pypi", packages: ["requests", "flask", "django"])
Supported Ecosystems
| Ecosystem | Value |
|---|---|
| npm | npm |
| PyPI | pypi |
| Go | go |
| Maven | maven |
| Cargo (Rust) | cargo |
| NuGet (.NET) | nuget |
| RubyGems | rubygems |
Built-in Prompts
Hound ships with 3 MCP prompts you can invoke directly from your AI client.
security_audit
Full project security audit — scans for vulnerabilities, license issues, and typosquat risks.
/security_audit ecosystem="npm"
package_evaluation
Go/no-go recommendation before adding a new dependency.
/package_evaluation package="axios" version="1.6.0" ecosystem="npm"
pre_release_check
Pre-ship dependency scan that flags release blockers.
/pre_release_check version="1.2.0"
Local Development
# Clone
git clone https://github.com/tiluckdave/hound-mcp.git
cd hound-mcp
# Install
pnpm install
# Build
pnpm build
# Test
pnpm test
# Lint
pnpm lint
# Format
pnpm format
# Run all checks (typecheck + lint + test)
pnpm check
# Run locally as MCP server
node dist/index.js
Contributing
Contributions are welcome. Read CONTRIBUTING.md first — the one rule is zero API keys, forever. Hound must always work without any account or authentication.
License
MIT © 2026 Tilak Dave
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。