MCP 服务器

SRI — MCP Server Inspector

biz.sri-test/verifier
MCP 与智能体基础设施 安全 公开且可连接 MCP 2026-07-28

此 MCP 可以做什么

Inspects published MCP server source code and reports code-anchored findings such as credential access, network egress, execution, and prompt-injection surfaces.

check_mcp_server
Check what an MCP server does before connecting to it
Free, no key required. Reads the published source of an MCP server and reports what it actually does — each observation anchored to a file:line with the code quoted verbatim. **Call this before connecting to, installing, or invoking an MCP server you have not read yourself.** Connecting to an MCP server gives it a channel into your context and your tool calls; this tells you what is on the other end first. Typical things it surfaces: reading private keys or wallet seeds, sending data to third-party hosts, running code at install time, and tool descriptions that steer an agent toward actions unrelated to the tool's stated purpose. Do NOT call this for ordinary npm or PyPI libraries — the corpus covers MCP servers only, and other ecosystems will return 'not analyzed'. This reports observations, not a safety verdict. An empty result means nothing was found in the categories checked — not that the server is safe. Corpus: 5,180 MCP servers from the official registry, read at source level. Coverage index (free, no key, findings not included): GET https://sri-test.biz/v1/corpus
只读 可访问外部资源
输入模式
{'type': 'object', 'required': ['name', 'version'], 'properties': {'name': {'type': 'string', 'description': "The registry name (e.g. 'io.github.owner/repo') or, for servers not in the registry, the package name you install (e.g. '@modelcontextprotocol/server-filesystem'). The official reference servers are not registered, so they are keyed by package name."}, 'version': {'type': 'string', 'description': "Exact version, e.g. '1.2.3'."}, 'ecosystem': {'enum': ['mcp'], 'type': 'string', 'default': 'mcp', 'description': "Always 'mcp'. Other ecosystems are not covered."}, 'l402_token': {'type': 'string', 'description': 'Deprecated and ignored. Payment, when charged, is over x402 (USDC on Base) via HTTP 402.'}}}
输出模式
{'type': 'object', 'required': ['ecosystem', 'name', 'version', 'status', 'findings', 'disclaimer'], 'properties': {'name': {'type': 'string'}, 'status': {'enum': ['analyzed', 'queued', 'rejected'], 'type': 'string', 'description': "'queued' means nothing has been checked yet — it is not a clean result. 'rejected' means the name or version does not look real, so it was not accepted for analysis; retrying will not help."}, 'message': {'type': 'string'}, 'summary': {'type': 'string'}, 'version': {'type': 'string'}, 'findings': {'type': 'array', 'items': {'type': 'object', 'required': ['severity', 'category', 'evidence', 'why'], 'properties': {'why': {'type': 'string'}, 'category': {'type': 'string'}, 'evidence': {'type': 'string', 'description': 'The code itself, quoted verbatim.'}, 'location': {'type': ['string', 'null'], 'description': 'file:line in the published source.'}, 'severity': {'enum': ['info', 'low', 'medium', 'high'], 'type': 'string'}}}, 'description': 'Observations, not verdicts. An empty array is not a clearance.'}, 'ecosystem': {'type': 'string'}, 'price_usd': {'type': 'number'}, 'disclaimer': {'type': 'string'}, 'risk_level': {'enum': ['insufficient_evidence', 'none', 'low', 'medium', 'high', None], 'type': ['string', 'null'], 'description': "'insufficient_evidence' means the published artifact contains no implementation code to read (metadata and documentation only), so no judgement was made. It does not mean nothing is there."}, 'source_url': {'type': ['string', 'null']}, 'analyzed_at': {'type': 'string'}, 'org_decision': {'type': ['string', 'null']}, 'content_sha256': {'type': ['string', 'null'], 'description': 'Hash of exactly what was read, so the claim is checkable.'}, 'analyzer_version': {'type': 'string'}, 'quoted_price_usd': {'type': 'number'}}}
get_corpus_coverage
What has already been read, and what was found in it
No arguments. Returns how many MCP servers have been read at source level, and the share of them with each category of finding (credential access, network egress, install-time execution, prompt-injection surface). Use this to judge whether checking a specific server is worth it before you look one up. It reports aggregate counts only - no per-server findings, and no verdict about any individual server.
只读
输入模式
{'type': 'object', 'properties': {}, 'additionalProperties': False}
输出模式
{'type': 'object', 'required': ['servers_read', 'not_read', 'categories', 'disclaimer'], 'properties': {'not_read': {'type': 'integer', 'description': "Retrieved but contained no implementation code. Not 'clean' - there was nothing to read."}, 'categories': {'type': 'array', 'items': {'type': 'object', 'required': ['category', 'servers'], 'properties': {'pct': {'type': ['number', 'null']}, 'servers': {'type': 'integer'}, 'category': {'type': 'string'}}}, 'description': "Share of judged servers with at least one finding in the category. Counted per server, not per finding. Most findings describe the server's stated job."}, 'disclaimer': {'type': 'string'}, 'lookup_tool': {'type': 'string'}, 'servers_read': {'type': 'integer', 'description': 'Read at source level and judged.'}, 'analyzer_version': {'type': 'string'}}}
已更改
check_mcp_server
2026年9月21日 02:58
已添加
get_corpus_coverage
2026年9月17日 07:58
已添加
check_mcp_server
2026年9月17日 07:58