Runheed

com.runheed由 runheed.com 發布官方發布者v1.0.1更新於 Nov 6, 2026

Production errors your coding agent can read: search, stack traces, context, trends, resolve.

已驗證Streamable HTTP可網頁執行開發者工具安全與監控

概覽

AI 產生的概覽

Runheed 透過 MCP 提供工作區的正式環境錯誤群組，讓助理搜尋錯誤、查看堆疊與脈絡、檢視趨勢，並將錯誤標記為已解決或已忽略。

可以做什麼
  • 依文字、狀態與專案搜尋錯誤群組，或列出最近發生的錯誤
  • 完整讀取單一錯誤群組，包含堆疊框架、麵包屑與 agent 脈絡
  • 檢視 24 小時、7 天或 30 天的錯誤統計，以及專案或工作區的健康分數
  • 變更錯誤狀態：附註解標記為已解決、忽略，或將已忽略的錯誤改回未解決
適用情境
適合讓編碼助理除錯 Node.js 或 Python 應用程式的真實正式環境例外：找出出錯的錯誤、閱讀堆疊與脈絡、修正程式碼，再用提交資訊標記為已解決。
執行需求
遠端 Streamable HTTP 端點 runheed.com/api/mcp，不需本機套件。以 OAuth 登入，或在 Authorization 標頭帶 Bearer agent key（al_...，來自 Settings > Agent keys）驗證。需要一個有應用程式回報錯誤的 Runheed 工作區，以及付費方案（每工作區每月 29 美元起）。
安裝前請注意
需要 Runheed agent key 或 OAuth 登入，agent key 應視為機密。有三個工具會變更錯誤狀態（解決、忽略、取消忽略），但可改回。錯誤資料（含受影響使用者與工作階段脈絡）會從託管的 Runheed 服務讀取。方案依工作區按月計費。

安裝

在 SourceWeft 中

  1. 開啟 儀表板中的 Runheed，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。

交給代理程式安裝

把這段提示詞貼上到 Claude Code、Codex、Cursor 等代理程式，或 SourceWeft 對話中。代理程式會讀取該伺服器的安裝指南，先向你展示發布者、端點和登入方式，經你同意後再把它新增到你的客戶端。金鑰不會寫進任何命令。

閱讀 https://sourceweft.com/mcp/com.runheed%2Frunheed/install.md，並按其中說明把這個 MCP 伺服器接入我的客戶端。修改任何設定之前，先告訴我它的發布者、端點和登入方式，等我確認。不要把 API 金鑰寫進命令或我會提交的檔案；問我金鑰應從哪裡取得。

檢視代理程式所依據的安裝指南

其他 MCP 客戶端

把它新增到你客戶端的 mcpServers 設定中。

{
  "mcpServers": {
    "runheed": {
      "type": "http",
      "url": "https://runheed.com/api/mcp"
    }
  }
}

README

Runheed

Error tracking your coding agent can read. Runheed catches exceptions in your Node.js or Python app and serves them over MCP, so Claude Code, Cursor, Codex or any MCP client can read the stack trace, fix the bug and mark it resolved.

  • Website: https://runheed.com
  • MCP server: https://runheed.com/api/mcp (Streamable HTTP). Sign in with OAuth (claude.ai, ChatGPT, Claude Code, Cursor, Codex), or send Authorization: Bearer al_... with an agent key.
  • Setup for agents: https://runheed.com/SETUP.txt

This repo holds the Claude Code plugin, the Google Antigravity plugin, the agent skill and the MCP config. The Runheed service itself is hosted.

Install

Claude Code plugin (signs you in on first use, no key to copy):

/plugin marketplace add getrunheed/runheed-plugin/plugin install runheed@runheed

Or add the server directly:

ClientHow
Claude Codeclaude mcp add --transport http --scope user runheed https://runheed.com/api/mcp, then run /mcp and choose Authenticate
Codexcodex mcp add runheed --url https://runheed.com/api/mcp then codex mcp login runheed
claude.ai / Claude DesktopCustomize > Connectors > Add custom connector, URL https://runheed.com/api/mcp
ChatGPTPlugins > + > Add custom MCP server, URL https://runheed.com/api/mcp, OAuth
CursorAdd {"mcpServers": {"runheed": {"url": "https://runheed.com/api/mcp"}}} to ~/.cursor/mcp.json, then sign in from Cursor Settings > MCP
Google Antigravityagy plugin install https://github.com/getrunheed/runheed-plugin (this repo is also an Antigravity plugin: .antigravity-plugin/plugin.json + mcp_config.json), or add {"mcpServers": {"runheed": {"serverUrl": "https://runheed.com/api/mcp"}}} to ~/.gemini/config/mcp_config.json and click Authenticate under Customizations
Gemini CLIgemini extensions install https://github.com/getrunheed/runheed-plugin

Then add the SDK to your app (Node.js / TypeScript, Python, or plain HTTP). Your agent can do both steps from https://runheed.com/SETUP.txt.

Install the skill in any agent

skills/runheed/SKILL.md tells an agent when to reach for Runheed and how to set it up. It's the same file served at https://runheed.com/SKILL.md.

What your agent can do

ToolWhat it does
runheed_search_errorsSearches the workspace's error groups by text (type or message), status and project.
runheed_get_errorReturns one error group in full: message, status, counts, first and last seen, environment, agent context, the top stack frames and the latest breadcrumbs.
runheed_get_recentReturns the error groups seen most recently, newest first.
runheed_get_trendsReturns error statistics for the last 24 hours, 7 days or 30 days.
runheed_get_error_contextReturns everything recorded around an error: breadcrumbs, runtime, environment, release, the affected user and the agent session.
runheed_resolve_errorMarks an error group resolved, with an optional note such as the fixing commit.
runheed_ignore_errorMarks an error group ignored, for known or expected errors.
runheed_unignore_errorMoves an ignored error group back to unresolved.
runheed_get_project_healthReturns a 0 to 100 health score for a project, with recent issues and recommendations.

Read tools are read-only; the three status tools only flip an error's status, which can be flipped back.

Pricing

One flat price per workspace, every plan includes the MCP server:

  • Solo: $29 a month. 50,000 events a month, 3 projects, 30-day history.
  • Pro: $79 a month. 250,000 events a month, 15 projects, 90-day history.
  • Scale: $199 a month. 1,000,000 events a month, unlimited projects, 180-day history.

Events over your plan's limit are rejected, never charged. Details: https://runheed.com/pricing

Support and contributing

Questions: [email protected]. Issues and pull requests for the skill and plugin files are welcome here.

License

MIT. Runheed is made by Great Work LLC.

來源：README.md，提交 ab8079a

工具

9

runheed_search_errors

Search errors

Searches the workspace's error groups by text (type, message, route, file or extra data), status and project. Returns each matching group with its status, type, message, route, top stack frame, event count, last-seen time and ID.

輸入參數limitquerystatusprojectId

runheed_get_error

Get an error

Returns one error group in full: message, status, counts, first and last seen, environment, agent context, the top stack frames and the latest breadcrumbs.

輸入參數errorId必填

runheed_get_recent

Get recent errors

Returns the error groups seen most recently, newest first, with status, type, message, route, top stack frame, event count, last-seen time and ID.

輸入參數limitprojectId

runheed_get_error_context

Get error context

Returns everything recorded around an error: every breadcrumb, platform, runtime, OS, environment, release, the affected user, the agent session and extra data.

輸入參數errorId必填

runheed_resolve_error

Resolve an error

Marks an error group resolved, with an optional note such as the fixing commit. The group reopens on its own if the error happens again. Returns a confirmation.

輸入參數commenterrorId必填

runheed_ignore_error

Ignore an error

Marks an error group ignored, for known or expected errors, so new occurrences stop alerting. Returns a confirmation.

輸入參數reasonerrorId必填

runheed_unignore_error

Stop ignoring an error

Moves an ignored error group back to unresolved, so new occurrences alert again. Returns a confirmation.

輸入參數errorId必填

runheed_get_project_health

Get project health

Returns a 0 to 100 health score for a project (or the whole workspace), errors in the last 24 hours, the unresolved count, the error rate, recent issues and recommendations.

輸入參數projectId

版本歷史

2
  1. v1.0.1最新Nov 6, 2026
  2. v1.0.0Nov 3, 2026

評分

已滿足適用檢查項的 100%

該評分衡量此收錄條目的完整程度，依據註冊中心條目、端點與公開來源計算得出。它不是對伺服器本身的評價，也不表示伺服器安全或目前可用：可連通性是帶日期的檢查結果，評分也不包含安裝量或使用者評價。

  • 近 30 天內可連通必需25 分
  • 已宣告認證方式必需10 分
  • 已列出工具必需15 分
  • 描述必需15 分
  • 圖示10 分
  • 文件15 分
  • 授權條款10 分