菜单
Use whenever the task involves an external integration connected through Executor — Notion, Context7, Supabase, Vercel, or any other MCP surfaced by mcp__executor__execute. Ensures these are always accessed through Executor instead of a direct MCP server, the public API, or the CLI.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/astrxnomo/dotfiles --skill mcp-integrations命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Shared vocabulary for designing deep modules. Use when the user wants to design or improve a module's interface, find deepening opportunities, decide where a seam goes, make code more testable or AI-navigable, or when another skill needs the deep-module vocabulary. Spanish triggers: "arquitectura", "estructura del código", "cómo organizo", "dónde pongo", "refactor", "refactoriza", "acoplado", "separar responsabilidades", "capas", "módulo", "interfaz".
Diagnosis loop for hard bugs and performance regressions. Use when the user says "diagnose"/"debug this", or reports something broken/throwing/failing/slow. Spanish triggers: "no funciona", "está roto", "falla", "da error", "revienta", "tira error", "se cae", "no anda", "por qué no", "está lento", "diagnostica", "depura", "arregla el bug", "no me sale".
Build and sharpen a project's domain model — the CONTEXT.md glossary and the ADRs in docs/adr/. Use when the user wants to pin down domain terminology or a ubiquitous language, record an architectural decision, or when another skill needs to maintain the domain model. Spanish triggers: "glosario", "cómo llamamos a", "terminología", "nombres del dominio", "modelo de dominio", "documenta la decisión", "ADR", "deja constancia de por qué".
基于 SOC 职业分类
正在显示 SKILL.md
| name | mcp-integrations |
| description | Use whenever the task involves an external integration connected through Executor — Notion, Context7, Supabase, Vercel, or any other MCP surfaced by mcp__executor__execute. Ensures these are always accessed through Executor instead of a direct MCP server, the public API, or the CLI. |
All external integrations go through Executor (mcp__executor__execute), never a direct MCP server, the public API, or a CLI shortcut. Call skills({ name: "execute" }) inside Executor first if you're unsure how to write the sandboxed code.
Discover, don't memorize. Connections and tool names change over time. Use tools.search({ query: "...", namespace: "<integration>_mcp" }) (or the bare integration name, e.g. "vercel") to find the exact tool path and argument shape. Any MCP added to Executor shows up here automatically — this skill does not keep an inventory of them.
Several integrations have more than one connected account/organization, distinguished by a .user.<name> suffix on the namespace (e.g. notion_mcp.user.felipegiraldo, notion_mcp.user.centrodeprototipado). Which ones are split changes over time, so check with tools.search rather than assuming a given integration has only one.
felipegiraldo is the default; centrodeprototipado is only for that project.context7_mcp)Two-step flow, worth remembering because it isn't obvious from the tool list:
resolve_library_id({ libraryName, query }) — both required; query is the user's full question, improves relevance ranking.query_docs({ libraryId, query }) — use the library ID from step 1 (prefer exact name match, higher benchmark score, and version-specific IDs when the user names a version).Use it when the user asks about libraries, frameworks, or API references, or needs current code examples instead of relying on training data.