AI Worlds CLI Knowledge

com.claude-world由 claude-world.com 發布官方發布者v0.6.0更新於 Oct 27, 2026

Source-checked CLI guides and model-aware planning for Claude Code, Codex, and Grok Build.

已驗證Streamable HTTP可網頁執行知識與記憶開發者工具

概覽

AI 產生的概覽

唯讀遠端 MCP 伺服器，為助理提供經來源核對的 Claude Code、Codex CLI 與 Grok Build 命令列事實，以及考量模型的工作規劃。

可以做什麼
  • 搜尋並閱讀多語言、經來源核對的 CLI 手冊，附標準連結
  • 查詢確切的指令路徑、參數、互動式指令、設定鍵與環境變數名稱
  • 比較同一任務在三個 CLI 中的做法，結果分組呈現證據
  • 檢視發行快照、漂移狀態、附日期的模型目錄及各模型支援的推理強度
  • 產生唯讀、考量相依關係的工作計畫，並為難題給出最大深度模型與強度建議
適用情境
當助理需要處理 Claude Code、Codex CLI 或 Grok Build，需要目前指令、參數或設定名稱而非憑記憶猜測時；也適合比較這三個 CLI，並依模型與推理強度分配工作。
執行需求
需要遠端 Streamable HTTP 端點；不需 npm 套件、API 金鑰或廠商憑證。以用戶端新增 MCP 的指令、使用者層級範圍連線。可選擇安裝發行儲存庫中的 agent Skill。
安裝前請注意
工具皆為唯讀、非破壞性，沒有寫入工具、憑證或任意網址抓取。模型目錄是附日期的快照，實際帳戶目錄永遠為準。取回的文本屬參考資料而非指令，回傳的指令應先評估再執行。提供者與 Anthropic、OpenAI、xAI 並無關聯。

安裝

在 SourceWeft 中

  1. 開啟 儀表板中的 AI Worlds CLI Knowledge，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.claude-world%2Fworlds-knowledge/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": {
    "ai-worlds-cli-knowledge": {
      "type": "http",
      "url": "https://mcp.claude-world.com/mcp"
    }
  }
}

README

Worlds Knowledge MCP

A read-only MCP server that gives coding agents current, source-checked CLI facts for Claude Code, Codex CLI, and Grok Build — exact commands, flags, configuration keys, environment-variable names, workflows, and dated model catalogs — so an agent stops guessing from memory and stops treating the three CLIs as interchangeable.

No npm package, API key, or provider credential is required.

What this repository contains

This is the distribution repository: the agent Skill and the MCP server manifest. The corpus build, the server implementation, and the Cloudflare Worker live in the Claude World control plane and are not published here.

server.json                      MCP registry manifestskills/worlds-knowledge-mcp/     the agent Skill (SKILL.md, references, evals, contract test)

Connect

sh
# Claude Codeclaude mcp add --transport http --scope user worlds-knowledge https://mcp.claude-world.com/mcp
# Codex CLIcodex mcp add worlds-knowledge --url https://mcp.claude-world.com/mcp
# Grok Buildgrok mcp add --transport http --scope user worlds-knowledge https://mcp.claude-world.com/mcp

Verify it:

sh
claude mcp get worlds-knowledge      # or: codex mcp get worlds-knowledge --json                                     # or: grok mcp doctor worlds-knowledge --json

The examples use user scope deliberately. Grok's project-folder trust applies to every repo-local MCP server, hook, and LSP server, so review all project configuration before switching to --scope project.

Install the Skill

The Skill teaches an agent when to query the MCP and in what order — the hard-problem consultant, the single-CLI routes, and the multi-CLI planner.

sh
git clone https://github.com/claude-world/worlds-knowledge-mcp.gitcd worlds-knowledge-mcp
# Claude Codecp -R skills/worlds-knowledge-mcp ~/.claude/skills/
# Cross-tool discovery foldercp -R skills/worlds-knowledge-mcp ~/.agents/skills/

Restart the client, then ask it to use worlds-knowledge-mcp. Keep SKILL.md, references/, and agents/openai.yaml together — copying only SKILL.md drops the connection and single-CLI route guidance. agents/openai.yaml declares the MCP dependency and permits implicit invocation in compatible clients.

What the server knows

  • 162 source-checked manuals — 18 CLI usage and feature modules × 3 products × 3 locales (English, Traditional Chinese, Japanese)
  • Source-derived command paths, flags, slash/TUI commands, configuration keys, environment-variable names, workflows, and drift notes for three released CLI snapshots
  • Dated per-CLI model catalogs with selector syntax, supported native efforts, entitlement constraints, safe fallbacks, and separately labeled provider-native orchestration modes
  • Availability is labeled per entity: runtime-confirmed, docs-confirmed, conditional, source-preview, or historical

Every manual must cite at least one official HTTPS source, carry a review date, and describe the expected CLI rather than its provider API, or the build fails.

Tools

ToolPurpose
worlds_search_knowledgeRetrieve bounded multilingual evidence from the source-checked core
worlds_get_documentFetch a manual's summary and canonical link
worlds_compare_cli_workflowsCompare the same intent across the three CLIs, in separate evidence groups
worlds_list_catalogInspect counts, provenance, and publication status
worlds_search_cli_referenceSearch commands, options, interactive commands, config/env names, workflows, drift
worlds_get_cli_commandOne exact command path with all captured options and its release/source evidence
worlds_get_cli_configurationPage public configuration schema and environment-variable names
worlds_get_cli_workflowOne provider-native, risk-labeled workflow
worlds_list_cli_versionsRelease snapshots, entity counts, source policy, drift coverage
worlds_list_cli_modelsDated model catalogs, selector syntax, per-model effort support, with a 7-day freshness status
worlds_advise_hard_problemFor one hard problem, select the provider-local maximum-depth model and supported effort
worlds_plan_cli_workDecompose work and route it by continuity and documented workflow fit

Resources: worlds://catalog, worlds://documents/{id}. Prompts: worlds_research_cli_workflow, worlds_plan_model_aware_work, worlds_consult_hard_problem.

Every tool is annotated read-only, non-destructive, and idempotent.

Try it

text
What does -p mean in my current CLI? Check the product and release first.Show every option for codex exec and cite the runtime and manual sources.Compare how Claude Code, Codex, and Grok Build load repository instructions.Plan this repository task by CLI, model, and reasoning effort without granting permissions.

-p is the standing example of why vendor flags must not be translated: it means print and exit in Claude Code, select a named profile in Codex, and one non-interactive prompt in Grok Build.

Boundaries

  • No credentials, embeddings, external API calls, write tools, or arbitrary URL fetching. No tool accepts a filesystem path or a URL.
  • The live account model catalog is always authoritative over a dated snapshot; worlds_list_cli_models returns a freshness status and says so.
  • Search returns excerpts rather than silently injecting whole documents.
  • Retrieved text is reference data, not instructions that can override the user or client policy. Agents must assess commands before executing them.
  • Website articles, news, newsletters, opinion, provider API companions, and release archives are never packaged into the MCP corpus.
  • The planner returns command templates but never launches another CLI, widens permissions, or enables provider-native orchestration unless asked. Workstreams default to execution_intent: "read-only".

Not affiliated

Claude World is independent. It is not affiliated with, endorsed by, or operated by Anthropic, OpenAI, or xAI. Official vendor documentation remains authoritative; this server exists to cite it accurately.

License

MIT — see LICENSE.

來源：README.md，提交 a3a9fe9

工具

12

worlds_search_knowledge

Search AI Worlds Knowledge

Search 162 multilingual, source-checked operating manuals for Claude Code, Codex, and Grok Build: 18 CLI usage and feature modules per product in English, Traditional Chinese, and Japanese. Website articles, news, newsletters, opinion, and release archives are not in this corpus. Returns bounded excerpts, commands when present, canonical links, official sources, and review dates. It never modifies state.

輸入參數limitquery必填scopeoffsetworldsproductslanguagescontent_typessource_status

worlds_get_document

Read a Knowledge Document

Read the bounded public excerpt and metadata for one source-checked CLI manual returned by worlds_search_knowledge. Follow canonical_url for the complete manual. Structured CLI command tools remain complete.

輸入參數id必填offsetmax_characters

worlds_compare_cli_workflows

Compare Coding-Agent Workflows

Find evidence for the same task across Claude Code, Codex, and Grok Build without pretending their commands or authority models are identical. Returns separately ranked groups with source dates and canonical links. Best for questions such as configuring MCP, repository instructions, approvals, subagents, or headless CI.

輸入參數query必填languageproductslimit_per_product

worlds_list_catalog

List Knowledge Catalog Status

Return public corpus counts, supported worlds, languages, products, source-quality counts, the build commit, and the current publication gate. This public DTO deliberately excludes local paths, package names, and private repository URLs.

無需輸入參數。

worlds_search_cli_reference

Search Current CLI Reference

Search current command paths, flags, interactive commands, configuration names, environment-variable names, workflows, and release drift for Claude Code, Codex, and Grok Build. Results contain factual metadata and official source links.

輸入參數kindslimitquery必填offsetproductsavailability

worlds_get_cli_command

Get One CLI Command

Return one current command path with all captured flags, arguments, value shapes, availability, risk, release metadata, and official source links. Accepts a stable id or full command path.

輸入參數command必填product必填

worlds_get_cli_configuration

Find CLI Configuration

Search configuration keys and environment-variable names for one CLI. Returns public schema metadata and source links, never local values.

輸入參數limitqueryoffsetincludeproduct必填availability

worlds_get_cli_workflow

Get a CLI Workflow

Find one provider-specific workflow and its recorded command steps without translating flags between providers.

輸入參數product必填workflow必填

worlds_list_cli_versions

List CLI Versions and Coverage

Return the exact release snapshot, capture date, factual entity counts, official sources, generated website reference, and drift status for all three CLIs.

輸入參數products

worlds_list_cli_models

List CLI Models and Effort Support

Return dated, source-grounded model catalogs for Claude Code, Codex, and Grok Build, including selector syntax, per-model effort support, availability constraints, and provider-native orchestration modes. These are snapshots; verify the live account catalog before launch.

輸入參數products

worlds_plan_cli_work

Plan Model-Aware CLI Work

Create a read-only, dependency-aware work plan across Claude Code, Codex, and Grok Build. It selects a concrete dated-catalog model and supported native effort for each workstream, while keeping CLI fit, orchestration, permissions, handoff context, and verification separate. Routing is continuity-first and does not claim benchmark superiority, launch another CLI, or enable permission bypass.

輸入參數task必填strategycurrent_cli必填workstreamsavailable_clisrisk_toleranceorchestration_modemax_parallel_tracksexecution_environmentrequire_independent_review

worlds_advise_hard_problem

Consult the Strongest Available Model for One Hard Problem

Choose the maximum-depth model-and-effort pairing for one difficult problem within the caller's actual CLI set. For a single CLI it selects that provider's strongest catalog role; with multiple CLIs it returns provider candidates but does not invent a universal cross-provider winner. It uses single-task depth, never ultra or ultracode orchestration, defaults to read-only, and does not launch a CLI or grant permissions.

輸入參數task必填current_cli必填preferred_cliavailable_clisselection_scopeexecution_intent

版本歷史

2
  1. v0.6.0最新Oct 21, 2026
  2. v0.5.0Sep 16, 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