原型 · 预览版:收录官方注册中心里的全部服务器,每 30 分钟同步一次。评价和账户功能将在后续阶段推出。
PopNx MCP为使用 AI 的人准备的 MCP 服务器 · 用大白话讲清楚

Toolstem Sec MCP Server

作者:toolstem版本 0.1.9i★ –i在你的电脑上运行
注册名称io.github.toolstem/toolstem-sec-mcp-server

安装

概览

作者提供的说明(英文原文)

SEC filing MCP for AI agents — insider signals, 13F flow, material events. Pre-computed.

PopNx MCP 的三大支柱

这三项永远不会合并成一个分数:意见、测量和人气是不同的东西。

它能帮你做什么

我们还没有为它写通俗的介绍。下面是我们从注册中心读取到的事实。

  • 连接方式: 在你的电脑上运行
  • 传输方式: stdio
  • 分类: 其他

安装

选择你使用的 AI 应用,然后复制设置。

安装之前

  • 基础检查:1 项中通过 1 项
  • 在你的电脑上运行,拥有你的用户账户的权限(文件、网络、应用)。
  • 需要你自己填写: SEC_USER_AGENT_CONTACT

你需要

  • npm已安装 Node.js(用于 npx 命令)

你用的是哪个应用?

在上方选择你的应用。我们会记住它,下次按钮会显示“安装到 …”。

打开 Claude Desktop → Settings → Developer → Edit Config,把下面的内容粘贴到 claude_desktop_config.json

{
  "mcpServers": {
    "toolstem-sec": {
      "command": "npx",
      "args": [
        "-y",
        "toolstem-sec-mcp-server"
      ],
      "env": {
        "SEC_USER_AGENT_CONTACT": "your-value"
      }
    }
  }
}

在终端中运行这条命令

claude mcp add toolstem-sec --env SEC_USER_AGENT_CONTACT=your-value -- npx -y toolstem-sec-mcp-server

运行前请先填写空着的值。

在 Cursor 中打开

或者自己复制设置

粘贴到项目中的 .cursor/mcp.json,或 ~/.cursor/mcp.json

{
  "mcpServers": {
    "toolstem-sec": {
      "command": "npx",
      "args": [
        "-y",
        "toolstem-sec-mcp-server"
      ],
      "env": {
        "SEC_USER_AGENT_CONTACT": "your-value"
      }
    }
  }
}

浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。

在 VS Code 中打开

或者自己复制设置

粘贴到项目中的 .vscode/mcp.json

{
  "servers": {
    "toolstem-sec": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "toolstem-sec-mcp-server"
      ],
      "env": {
        "SEC_USER_AGENT_CONTACT": "your-value"
      }
    }
  }
}

浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。

这个服务器在你的电脑上运行,而 ChatGPT 只能连接互联网上的服务器。请选择其他应用。

你的选择通过 Cookie 保存在这台设备上,无需账户。

需要你填写的值

  • SEC_USER_AGENT_CONTACT env
    Contact email embedded in the SEC EDGAR User-Agent (required by SEC fair-access policy). Defaults to admin@toolstem.com.

这些设置是根据作者在注册中心的条目生成的。如果 README 的写法不同,请以 README 为准。

它能做什么

它在你的电脑上运行;安装后,工具列表会显示在你的 AI 应用中。

基础检查i

我们检查的是任何人都能看到的信息;我们没有阅读代码。不适用的检查项不计入总数。 · 检查于 2026年11月6日

  • ✓ 通过

    作者与仓库一致

    注册名称已通过 GitHub 账户 toolstem 验证,该账户拥有此仓库

  • … 尚未检查

    仍在维护

    我们还没有从 GitHub 读取这个仓库

  • … 尚未检查

    有许可证

    我们还没有从 GitHub 读取这个仓库

  • – 不适用

    端点有响应

    它在你的电脑上运行,没有互联网端点

  • – 不适用

    使用 OAuth 登录

    在你的电脑上运行,改用本地设置

  • … 尚未检查

    工具描述中没有隐藏指令

    它在你的电脑上运行;我们只能读取互联网端点的工具列表

与“其他”中的其他服务器比较i

  • 星标数多于此分类中其他 32415 个服务器的 0%
  • 发布到注册中心的时间比此分类中其他 32415 个服务器的 13% 更新
  • 基础检查结果等于或优于此分类中其他 32415 个服务器的 100%

与注册中心里这个分类中所有活跃的服务器比较。只有当发布者拥有该仓库时,星标数才计入。

常见问题

Toolstem Sec MCP Server 安全吗?

它在 1 项适用的基础检查中通过了 1 项。我们没有阅读代码。在添加密钥或令牌之前,请先看看这个服务器要求哪些权限,并使用只具备所需权限的密钥。

它需要哪些密钥或代码?

注册中心没有列出必需的密钥,不过有些服务器会在首次使用时要求你登录。

它在我的电脑上运行还是在云端?

它在你的电脑上运行。你的 AI 应用会在需要时启动它,数据不会经过作者的服务器(除非它自己调用在线服务)。

它还在维护吗?

GitHub 仓库最后更新于–,最新版本(0.1.9)于5个月前进入注册中心。

哪些应用可以使用它?

支持 MCP 的应用,例如 Claude Desktop、Claude Code、Cursor 和 VS Code。你的电脑上还需要有 Node.js。

有什么可以替代它?

同一分类中有:Spec Driven Development, Remote Desktop Commander, Safe Fetch

替代选择

同一分类中其他作者的服务器,按星标数排序。

如何看卡片:★ GitHub 星标i三大支柱:评价 · 检查 · 使用情况i更新 = 最新版本发布到注册中心的时间i

Spec Driven Development

作者:formulahendry

MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.

在你的电脑上运行★ 440MIT14个月前更新
安装

Remote Desktop Commander

作者:desktop-commander

Remote Desktop Commander: use files, terminal and processes on your own computers from any AI app.

在云端运行★ 211昨天更新
安装

Safe Fetch

作者:Sanoy24

MCP server for fetching URLs, safe against SSRF, DNS rebinding, and redirect-to-internal attacks.

在你的电脑上运行今天更新
安装

作者的 README

打开 README

按原文从作者的仓库复制。PopNx MCP 没有审查或背书其内容。HTML、图片和脚本已被移除。链接会离开本站。

SEC EDGAR MCP Server — Insider Signals, 13D Activist Risk & Filing Intelligence

SEC EDGAR intelligence for AI agents. Five composite tools that pre-compute high-value signals directly from SEC EDGAR's public submissions API, returned as structured JSON.

No SEC API key required. Data is sourced directly from SEC EDGAR's public submissions API. A built-in sliding-window rate limiter keeps traffic under SEC's 10 rps fair-access ceiling automatically.


Quickstart — hosted endpoint (recommended)

Point your MCP client or agent at the hosted endpoint. No API key, no infra, no setup. Billing is per-call via x402 — the agent's wallet pays directly in USDC on Base mainnet.

https://mcp.toolstem.com/mcp/sec
  • No SEC API key, no signup, no marketplace account — the agent's wallet pays directly.
  • No infrastructure — nothing to install, host, or keep running.
  • No setup — connect an MCP client and call a tool.
  • initialize and tools/list are free (discovery and schema introspection).
  • tools/call is tiered per tool (see Pricing below).
Claude Desktop

Drop this into your claude_desktop_config.json:

{

"mcpServers": { "toolstem-sec": { "command": "npx", "args": [ "-y", "mcp-remote", "https://mcp.toolstem.com/mcp/sec" ] } } }

Restart Claude Desktop, then ask: *"Has TSLA disclosed any material 8-K events in the last 90 days?"*

Any MCP client (LangChain.js)

The official @langchain/mcp-adapters library connects directly to the hosted URL:

import { MultiServerMCPClient } from "@langchain/mcp-adapters";

import { ChatOpenAI } from "@langchain/openai"; import { createReactAgent } from "@langchain/langgraph/prebuilt";

const client = new MultiServerMCPClient({ toolstem_sec: { transport: "http", url: "https://mcp.toolstem.com/mcp/sec", // Add your x402-signing middleware via headers, OR run an x402 // proxy locally and point url at it. See https://www.x402.org/clients. }, });

const tools = await client.getTools(); const agent = createReactAgent({ llm: new ChatOpenAI({ model: "gpt-4o-mini" }), tools }); await agent.invoke({ messages: "Has TSLA disclosed any material 8-K events in the last 90 days?" });

LangChain quick-start (langchain-toolstem)

The langchain-toolstem wrapper handles x402 payment for you — pass a funded wallet key and the SEC tools are included automatically:

import { createToolstemTools } from 'langchain-toolstem';

const tools = await createToolstemTools({ walletPrivateKey: process.env.WALLET_KEY }); // SEC tools included automatically — agents pay per call in USDC

Prefer to run the server yourself over stdio/HTTP? See Advanced: self-host at the bottom.

Try the tools live in the Toolstem playground.


Pricing

  • MCP `initialize` and `tools/list` are free — discover the server and its tool surface without paying anything.
  • `tools/call` is tiered per tool, paid in USDC on Base mainnet via x402. No API key, no signup, no marketplace account — agents pay directly from their own wallet.
TierPrice per callTools
Cheap$0.005get_company_filings_summary
Standard$0.05get_insider_signal, get_institutional_signal
Premium$0.50get_material_events_digest, compare_disclosure_signals

Per-tool breakdown:

ToolTierPer call
get_company_filings_summaryStandard$0.005 USDC
get_insider_signalStandard$0.05 USDC
get_institutional_signalStandard$0.05 USDC
get_material_events_digestPremium$0.50 USDC
compare_disclosure_signalsPremium$0.50 USDC
How billing works

Toolstem uses the x402 payment protocol. Agents pay per call in USDC on Base — no API keys, no subscriptions, no invoices. The agent's wallet settles each call automatically via EIP-3009.

See the live pricing page on toolstem.com/sec/ for current rates.


The five tools

All five tools are composite/curated (they compute derived signals or aggregate across multiple EDGAR endpoints — no raw passthroughs). Annotations: readOnlyHint: true, destructiveHint: false, idempotentHint: true, openWorldHint: true.

#ToolRequired inputOptional input (default)Tier (price/call)
1get_company_filings_summaryticker_or_cik (string)—Cheap ($0.005)
2get_insider_signalticker_or_cik (string)lookback_days (int 1–730, default 90)Standard ($0.05)
3get_institutional_signalticker_or_cik (string)quarters_back (int 1–20, default 4)Standard ($0.05)
4get_material_events_digestticker_or_cik (string)lookback_days (int 1–1825, default 365)Premium ($0.50)
5compare_disclosure_signalstickers_or_ciks (string[2..5])—Premium ($0.50)

1. get_company_filings_summary

Overview of a company's filing activity: last 20 filings + computed signals.

SignalDescription
filing_velocityACCELERATING / NORMAL / SLOWING vs. trailing 365-day average
material_event_count_90dCount of 8-K filings in the last 90 days
disclosure_volume_trendRISING / STABLE / FALLING based on 10-K size comparison
latest_form_typesUnique form types filed in the last 90 days

Example output (abbreviated):

{

"ticker": "AAPL", "cik": "0000320193", "company_name": "Apple Inc.", "signals": { "filing_velocity": "NORMAL", "material_event_count_90d": 4, "disclosure_volume_trend": "RISING", "latest_form_types": ["8-K", "4", "DEF 14A"] }, "meta": { "source": "sec_edgar_direct", "data_delay": "live" } }


2. get_insider_signal

Probes Form 3 / 4 / 4/A insider filing activity within a configurable lookback window. Required: ticker_or_cik. Optional: lookback_days (1–730, default 90).

Returns: recent_insider_filings[] (accession numbers + SEC URLs), net_transaction_count, buy_count, sell_count, and insider_signal.

v0.1 limitation — counts only. v0.1 returns counts and Form 4 references only; direction-aware buy/sell signals ship in v0.2 (Form 4 XML parsing). Today, insider_signal is null when filings exist in the window (direction unknown) and "NEUTRAL" when no insider filings exist (verified absence). buy_count / sell_count are 0 in v0.1.

Example output (abbreviated):

{

"ticker": "MSFT", "cik": "0000789019", "company_name": "MICROSOFT CORP", "lookback_days": 90, "insider_signal": null, "net_transaction_count": 0, "buy_count": 0, "sell_count": 0, "recent_insider_filings": [ { "accession_number": "0001127602-26-001234", "filing_date": "2026-04-15", "sec_url": "https://www.sec.gov/Archives/edgar/data/789019/000112760226001234/0001127602-26-001234-index.htm" } ], "meta": { "source": "sec_edgar_direct", "data_delay": "live" } }


3. get_institutional_signal

Probes for activist investor activity via SC 13D / 13D/A filings. Required: ticker_or_cik. Optional: quarters_back (1–20, default 4 ≈ 1 year).

FieldDescription
activist_risk_flagtrue if any SC 13D or 13D/A was filed in the last 365 days
recent_13d_filingsList of 13D filings with form type, date, and SEC URL

v0.1 limitation — activist flag only. v0.1 ships the live `activist_risk_flag` (from 13D/13D-A) and a list of 13D filings. Quarterly 13F XBRL parsing — which produces institutional_signal (ACCUMULATING / HOLDING / DISTRIBUTING) and recent_13f_count — ships in v0.2. Today those two fields are null / 0.

Example output (abbreviated):

{

"ticker": "NVDA", "cik": "0001045810", "company_name": "NVIDIA CORP", "quarters_back": 4, "institutional_signal": null, "recent_13f_count": 0, "activist_risk_flag": false, "recent_13d_filings": [], "meta": { "source": "sec_edgar_direct", "data_delay": "live" } }


4. get_material_events_digest ⚡ Premium tier

Premium tier — $0.50 USDC per call on Base mainnet, settled via x402. See the live pricing page on toolstem.com/sec/ for current rates.

Severity-ranked digest of all 8-K and 8-K/A filings within a configurable lookback window. Each item code is mapped to a plain-English label and severity rating. Required: ticker_or_cik. Optional: lookback_days (1–1825, default 365).

SeverityExamples
🔴 REDCybersecurity incident (1.05), restatement (4.02), bankruptcy (1.03), delisting (3.01)
🟡 YELLOWAcquisition (2.01), new debt (2.03), executive departure (5.02)
🟢 GREENEarnings release (2.02), Reg FD (7.01), shareholder vote (5.07)

Returns: events[] (sorted newest-first), redflag_count, category_counts.

Example output (abbreviated):

{

"ticker": "TSLA", "cik": "0001318605", "company_name": "Tesla, Inc.", "lookback_days": 180, "redflag_count": 1, "category_counts": { "RED": 1, "YELLOW": 3, "GREEN": 7 }, "events": [ { "accession_number": "0001628280-26-005678", "filing_date": "2026-04-10", "form": "8-K", "items": [ { "code": "4.02", "label": "Non-Reliance on Previously Issued Financial Statements", "category": "financial", "severity": "RED" } ], "sec_url": "https://www.sec.gov/Archives/edgar/data/1318605/000162828026005678/0001628280-26-005678-index.htm" } ], "meta": { "source": "sec_edgar_direct", "data_delay": "live" } }


5. compare_disclosure_signals

Side-by-side comparison of 2–5 companies across all key disclosure signals. Required: tickers_or_ciks (string[2..5]). All lookups run in parallel.

Returns per-company: filing_velocity, material_event_count_90d, redflag_count_365d, activist_risk_flag, last_filing_date.

Returns winners (as CIKs, not tickers — cross-reference with the companies[] array): quietest_disclosure, most_active, most_redflags, activist_targets.

Example output (abbreviated):

{

"companies": [ { "ticker": "AAPL", "cik": "0000320193", "filing_velocity": "NORMAL", "material_event_count_90d": 4, "redflag_count_365d": 0, "activist_risk_flag": false, "last_filing_date": "2026-04-25" }, { "ticker": "MSFT", "cik": "0000789019", "filing_velocity": "ACCELERATING", "material_event_count_90d": 7, "redflag_count_365d": 0, "activist_risk_flag": false, "last_filing_date": "2026-04-26" } ], "winners": { "quietest_disclosure": "0000320193", "most_active": "0000789019", "most_redflags": null, "activist_targets": [] }, "meta": { "source": "sec_edgar_direct", "data_delay": "live" } }


Advanced: self-host

Most users should use the [hosted endpoint](#quickstart--hosted-endpoint-recommended) above — it needs no API key, no infrastructure, and no setup. This section is for users who specifically want to run the server themselves. When self-hosting you are responsible for running th

只显示了开头部分;其余内容请在仓库中阅读

问 PopNx