{
  "markdown": "# AgentGuard — security checks for AI agents (MCP server)\n\n<!-- mcp-name: io.github.shuaicongxiaomai/agentguard -->\n\nA free, open-source, single-binary [MCP](https://modelcontextprotocol.io) server that gives any\nAI agent (Claude Desktop, Cursor, Claude Code, …) security checks **in the loop**:\n\n| Tool | What it does | Data source |\n|---|---|---|\n| `scan_secrets` | Flags hard-coded secrets (API keys, tokens, private keys) in text/code/diffs before the agent commits, logs, or sends them. Runs locally; values are masked. | local (regex + entropy) |\n| `check_cve` | Whether a specific package version has known CVEs, with severity + fixed version. | [OSV.dev](https://osv.dev) (free) |\n| `scan_dependencies` | Scans a `go.mod` / `package.json` / `requirements.txt` and reports every vulnerable pinned dependency. | OSV.dev (free) |\n\nNo API key, no account, no telemetry. Everything runs locally except CVE lookups, which hit the\nfree public [OSV.dev](https://osv.dev) API.\n\n## Install\n\nPick whichever fits your setup. All three give the same stdio MCP server.\n\n### Docker (no Go toolchain needed)\n\n```sh\ndocker run -i --rm ghcr.io/shuaicongxiaomai/agentguard:latest demo\n```\n\nMCP client config (`mcpServers` block in Claude Desktop / Cursor / Claude Code):\n\n```jsonc\n{\n  \"mcpServers\": {\n    \"agentguard\": {\n      \"command\": \"docker\",\n      \"args\": [\"run\", \"-i\", \"--rm\", \"ghcr.io/shuaicongxiaomai/agentguard:latest\"]\n    }\n  }\n}\n```\n\n### Go install\n\n```sh\ngo install github.com/shuaicongxiaomai/agentguard@latest\n```\n\n```jsonc\n{ \"mcpServers\": { \"agentguard\": { \"command\": \"agentguard\" } } }\n```\n\n### Prebuilt binary\n\nDownload the archive for your OS/arch from the [Releases](https://github.com/shuaicongxiaomai/agentguard/releases)\npage, extract it, and point the config `command` at the absolute path to the binary.\n\n## Build from source\n\n```sh\ngo build -o agentguard .          # append .exe on Windows\n```\n\n## Three ways to run it\n\n**1. `demo` — see it work, zero config:**\n\n```sh\nagentguard demo\n```\n\nRuns all three tools on sample inputs and prints the results.\n\n**2. stdio (default) — local use; the MCP client launches the binary.** No port;\nthe client manages its lifecycle. Use any of the configs above, then ask your agent:\n*\"scan this for secrets\"*, *\"is lodash 4.17.20 vulnerable?\"*,\n*\"check my package.json for vulnerable deps.\"*\n\n**3. `serve` — Streamable HTTP service on a port (for a hosted/remote server):**\n\n```sh\nagentguard serve :8080          # or set PORT=8080\n# health:   GET  http://host:8080/healthz\n# MCP URL:  POST http://host:8080/mcp   (clients connect to this URL)\n```\n\nUse stdio for local use; use `serve` when you want a hosted server users connect to by URL\nwithout installing anything.\n\n## Notes\n\n- `scan_secrets` runs locally and **never returns raw secret values — only masked previews.**\n- `scan_dependencies` caps at 200 deps per call and queries OSV concurrently.\n- `check_cve` / `scan_dependencies` need network access to OSV.dev; `scan_secrets` is fully offline.\n\n## License\n\n[MIT](LICENSE).\n",
  "bytes": 3035,
  "sha": "e31fffaced78133ac8d996df1ba804d3b02535bf48eb72ace30a28165b31d31c",
  "repo_slug": "shuaicongxiaomai/agentguard",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_io_github_shuaicongxiaomai_agentguard_9344bb1c/readme"
}