Heropedia

io.github.SiliconRoshiBillv0.1.3更新於 Oct 10, 2026

Consult 324+ expert personas (Buffett, Jobs, Munger, Torvalds, ...) via 4 read-only registry tools.

已驗證Streamable HTTP可網頁執行Knowledge & MemoryProductivity & Workflow

概覽

AI 產生的概覽

讓助理從 Heropedia 取得權威的專家角色提示詞,並以該專家的視角作答。

功能
Heropedia 透過公開的 HTTP MCP 端點提供四個唯讀工具:getList 回傳分頁的「人物 × 角色」項目目錄,getListByHero 與 getListByRole 依名稱篩選(使用安全的正規表示式語法),getDetail 則回傳某一項目的完整權威 Markdown 提示詞。列表工具只回傳輕量中繼資料,例如 id、人物名稱、角色名稱與描述;只有 getDetail 會回傳完整角色提示詞。隨附的技能會運用這些工具,以指定專家的身分作答,可單輪完成,也可進行結構化的多輪「辦公時間」工作階段。
適用情境
當你希望以某位知名專家的視角取得第二意見時使用,例如審閱定價頁、到達頁、go/no-go 決策或程式重構。它既適合快速單輪諮詢,也適合較長的结构化工作階段——後者會提出追問,並給出診斷與一週行動計畫。
執行需求
需要一個位於 heropedia.org 的遠端 MCP 端點,使用 Streamable HTTP;不需要安裝套件、API 金鑰或帳號。需要網路連線,且該公開端點有 IP 流量限制(列表工具每分鐘 300 次請求,詳情每分鐘 60 次)。部分用戶端需要新增連接器或設定項目,辦公時間技能會在使用者家目錄下寫入本機筆記。
安裝前請注意
該端點公開且無需驗證,因此你在諮詢中貼上的內容會傳送到第三方伺服器。辦公時間技能會在使用者家目錄的某個目錄中寫入本機筆記。角色輸出是署名真實人物的生成文字,不應視為該人物的真實觀點或建議。

安裝

在 SourceWeft 中

  1. 開啟 儀表板中的 Heropedia,將其新增到工作區。
  2. 為需要使用其工具的對話啟用該服務。

Web executable,透過 Streamable HTTP。 遠端服務在工作區中設定後即可從網頁執行環境執行。

其他 MCP 客戶端

把它新增到你客戶端的 mcpServers 設定中。

{
  "mcpServers": {
    "heropedia": {
      "type": "http",
      "url": "https://www.heropedia.org/mcp"
    }
  }
}

README

Heropedia Plugin for Claude Code

Consult 324+ hand-crafted expert personas — Warren Buffett, Steve Jobs, Charlie Munger, Linus Torvalds, and more — directly from any Claude Code session.

This plugin bundles:

  • heropedia skill — triggers on phrases like "ask to...", "consult ", "channel ". Fetches the canonical persona prompt from Heropedia MCP before answering, so the reply is truly through that expert's lens (not an impression of it).
  • .mcp.json — registers https://www.heropedia.org/mcp as an MCP server so Claude Code can call getList, getListByHero, getListByRole, and getDetail.

The plugin does not ship any data. All 324+ personas live at heropedia.org and are fetched on demand.

Skills

This plugin bundles two skills that share the same MCP server:

  • heropedia — single-turn "ask an expert" lens. Triggers on phrases like "ask Warren Buffett to look at my deck", "as Steve Jobs, review this page", "consult Charlie Munger". Fetches the canonical persona prompt and replies in one turn through that lens. Best for a quick second opinion.
  • heropedia-office-hour — five-phase structured multi-turn session hosted by any hero. Triggers on "let hold office hours", "office hour with ", "office hour on ". The hero opens, asks 5–6 forcing questions one at a time (from a curated ## Office Hour Questions section if the persona has one, otherwise distilled on the fly), takes a direct position on your real problem, then gives you a diagnosis, a one-week action, and a red-flag warning. Saves local notes to ~/.heropedia/office-hours/. Best when you want to be pushed, not just answered.

Both skills refuse to invent personas — they always fetch canonical prompts from heropedia before speaking.

Install

Pick your AI. Every path targets the same endpoint (https://www.heropedia.org/mcp).

Anthropic — Claude Code (recommended)

bash
claude plugin marketplace add https://github.com/SiliconRoshiBill/heropedia-pluginclaude plugin install heropedia@heropedia

Auto-registers the MCP server as plugin:heropedia:heropedia. Open a new session and you're wired up.

Anthropic — Claude Desktop

Add to claude_desktop_config.json:

json
{  "mcpServers": {    "heropedia": {      "command": "npx",      "args": ["-y", "mcp-remote", "https://www.heropedia.org/mcp"]    }  }}

OpenAI — ChatGPT (web / macOS / Windows apps)

Available on Plus, Pro, Business, Enterprise, and Edu plans (not Free / Team).

  1. Settings → Connectors → Advanced → toggle Developer mode on.
  2. Settings → Connectors → Add custom connector.
  3. Paste MCP URL: https://www.heropedia.org/mcp
  4. Auth: No authentication. Save.
  5. Enable the Heropedia connector for chats where you want it.

Business/Enterprise workspaces: a workspace admin adds the connector under Workspace Settings → Connectors → New app.

OpenAI — Codex (CLI)

bash
curl -sSL https://raw.githubusercontent.com/SiliconRoshiBill/heropedia-plugin/main/codex/install-codex.sh | bash

Add --with-agents to also install the workflow-enforcing AGENTS.md:

bash
curl -sSL https://raw.githubusercontent.com/SiliconRoshiBill/heropedia-plugin/main/codex/install-codex.sh | bash -s -- --with-agents

Or manually add to ~/.codex/config.toml:

toml
[mcp_servers.heropedia]url = "https://www.heropedia.org/mcp"

OpenAI — Agents SDK

Python (openai-agents-python):

python
from agents import Agent, Runnerfrom agents.mcp import MCPServerStreamableHttp
async with MCPServerStreamableHttp(    params={"url": "https://www.heropedia.org/mcp"},    name="heropedia",) as server:    agent = Agent(        name="ConsultantAgent",        instructions=(            "When the user asks to consult a specific expert, call "            "getListByHero to find the entry, then getDetail to fetch "            "the canonical prompt. Apply that persona — do not paraphrase."        ),        mcp_servers=[server],    )    result = await Runner.run(agent, "Ask Warren Buffett to review my pricing page.")    print(result.final_output)

TypeScript (@openai/agents):

ts
import { Agent, Runner } from "@openai/agents";import { MCPServerStreamableHttp } from "@openai/agents/mcp";
const heropedia = new MCPServerStreamableHttp({  url: "https://www.heropedia.org/mcp",  name: "heropedia",});
const agent = new Agent({  name: "ConsultantAgent",  instructions:    "When the user asks to consult a specific expert, call getListByHero " +    "to find the entry, then getDetail to fetch the canonical prompt. " +    "Apply that persona — do not paraphrase.",  mcpServers: [heropedia],});
const result = await Runner.run(agent, "Ask Warren Buffett to review my pricing page.");console.log(result.finalOutput);

Google — Gemini CLI

This repo ships a gemini-extension.json. Follow the Gemini CLI extensions guide to add it, or paste this into your Gemini config:

json
{  "mcpServers": {    "heropedia": {      "httpUrl": "https://www.heropedia.org/mcp"    }  }}

Cursor

Add to your Cursor MCP config file:

json
{  "mcpServers": {    "heropedia": {      "url": "https://www.heropedia.org/mcp"    }  }}

Any HTTP MCP client

Direct JSON-RPC 2.0 POST to https://www.heropedia.org/mcp. See heropedia.org/mcp-guide for examples.

Try it

Once installed, just ask:

"Ask Warren Buffett to review my SaaS pricing page." "As Steve Jobs, tear apart this landing page." "Consult Charlie Munger on this go/no-go decision." "What would Linus Torvalds say about this refactor?" "Get a product critic to look at this feature spec."

Claude will:

  1. Recognize the persona intent.
  2. Call mcp__heropedia__getListByHero to find the entry.
  3. Call mcp__heropedia__getDetail to fetch the canonical prompt.
  4. Answer through that persona's lens.

If a name maps to multiple roles (e.g. Warren Buffett as investor vs. business analyst), Claude asks which lens fits your task before answering.

What's inside

heropedia-plugin/├── .claude-plugin/│   ├── plugin.json          # plugin manifest│   └── marketplace.json     # marketplace listing (installable via `claude plugin marketplace add`)├── .mcp.json                # MCP server registration├── skills/│   ├── heropedia/│   │   └── SKILL.md         # single-turn "ask an expert" skill│   └── heropedia-office-hour/│       └── SKILL.md         # five-phase multi-turn office-hour skill├── scripts/│   ├── validate-skill.sh    # structural lint for office-hour SKILL.md│   └── test-scenarios.md    # manual QA scenarios├── codex/                   # OpenAI Codex CLI installer + AGENTS.md├── gemini-extension.json    # Gemini CLI extension descriptor└── README.md

Requirements

  • Claude Code (any recent version).
  • Internet access — the MCP server is hosted at https://www.heropedia.org/mcp.
  • Nothing else. No API key. No account. Public endpoint, IP-rate-limited (300 req/min list, 60 req/min detail).

The MCP endpoint

Public HTTP MCP server at https://www.heropedia.org/mcp speaking MCP 2025-06-18 Streamable HTTP. Four read-only tools:

ToolPurpose
getListPaginated catalog of every hero × role entry
getListByHeroFilter by hero name (regex, case-insensitive, safe grammar)
getListByRoleFilter by role name (same regex rules)
getDetailFull canonical markdown for one entry

Returns lightweight metadata ({id, hero_name, role_name, description}) from LIST tools; only getDetail returns the full persona prompt.

User guide with install snippets, examples, and FAQ: heropedia.org/mcp-guide

Uninstall

bash
claude plugin uninstall heropedia

License

MIT. See LICENSE.

The persona content at heropedia.org is separately MIT-licensed in the SiliconRoshiBill/heropedia content repo.

Contribute a new hero

Anyone with a GitHub account can. Submit at heropedia.org/submit. The community votes on the best versions.

來源:README.md,提交 b8e962e

工具

0
工具後設資料尚未被收錄。

版本歷史

1
  1. v0.1.3最新Sep 16, 2026