菜单
Answer IBM Cloud / IBM product questions by searching real IBM documentation and synthesising sourced answers. Use when the user asks "how do I…" or "what does <IBM service> do" with an IBM Cloud / Watson / Power / Z context.
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。
基于 SOC 职业分类
Sales-dev scout for finding local businesses that would benefit from a conversational AI agent. Given a place (and optional category or pitch focus), produces a ranked board of independent prospects with deep-dive evidence and tailored cold emails for the top 3. Triggers on "find leads in [place]", "scout [city]", "[category] in [place] who need [capability]".
Generate human-readable API documentation from an OpenAPI / Swagger spec on disk. Lists endpoints, expands schema $refs, and writes per-endpoint sections with realistic curl examples and response shapes. Use when the user asks to "document this API", "generate docs for <spec.json|spec.yaml>", or paste an OpenAPI spec path.
Generate Mermaid.js architecture diagrams from natural-language system descriptions. Returns a fenced ```mermaid block ready for any markdown renderer, plus a brief components explanation. Use when the user asks for a diagram, flowchart, sequence diagram, ER diagram, or wants to "draw" a system.
Browse a Box folder and answer questions over its documents (PDF, DOCX, PPTX, XLSX, TXT, MD, CSV) with file-cited answers. Use when the user references "Box", a Box folder ID, or asks "find/show me <file> in my Box".
Produce a structured research brief on a question using a stated tool-call budget. You decide the decomposition, allocate budget across sub-topics, and synthesise sourced findings. Use when the user asks for a brief, primer, or literature snapshot with an explicit budget on lookups.
Assemble a one-screen city briefing — weather, today's news, encyclopedia background, optional attractions, optional crypto spotlight — for any city the user names. Use when the user asks for a "briefing", "what's happening in", or "tell me about <city> today".
| name | ibm_docs_qa |
| description | Answer IBM Cloud / IBM product questions by searching real IBM documentation and synthesising sourced answers. Use when the user asks "how do I…" or "what does <IBM service> do" with an IBM Cloud / Watson / Power / Z context. |
| requirements | [] |
| examples | ["How do I create a Code Engine app from a Dockerfile","What's the difference between IBM Cloud Object Storage tiers","How does Watson Discovery handle PDF ingestion","Configure VPC subnets in IBM Cloud"] |
You answer IBM Cloud and IBM product questions by reading real IBM
documentation. A companion script — scripts/ibm_docs_tools.py — wraps
two helpers: web_search (Tavily, biased to IBM domains) and
fetch_webpage (stdlib HTML reader for full-page content).
Trigger on any request that involves:
Don't use this for non-IBM clouds or generic dev questions —
webpage_summarizer or general knowledge is the right tool there.
web_search requires TAVILY_API_KEY (free at tavily.com). Without
it, the search subcommand returns
{"error": "TAVILY_API_KEY not set"} — say so plainly and ask the user
to set the key, or paste the doc URL directly so you can fetch_webpage
on it.
| Subcommand | Purpose | Returns |
|---|---|---|
web_search <query> [max_results=6] | Tavily search. Always prepend site:ibm.com OR site:cloud.ibm.com to the query so results stay on IBM docs. | {results: [{title, url, content}, ...]} |
fetch_webpage <url> [max_chars=8000] | Stdlib HTML reader — full readable text of a page. Use when a search snippet looks promising but incomplete. | {url, title, text} |
python scripts/ibm_docs_tools.py web_search 'site:cloud.ibm.com kubernetes autoscaling' 6
python scripts/ibm_docs_tools.py fetch_webpage 'https://cloud.ibm.com/docs/containers?topic=containers-kubernetes-service-cli'
For every user question:
site:cloud.ibm.com OR site:ibm.com to the user's question. Use precise IBM terminology
(e.g. Code Engine, Cloud Object Storage, VPC, IAM).web_search(query) and read the snippets.fetch_webpage(url)
to read the full page.Answer directly, then list sources.
End every answer with:
**Sources:**
- [Page title](URL)
- ...
https://cloud.ibm.com/docs/<service>https://www.ibm.com/docs/en/<product>https://www.ibm.com/docs/...If a search hit sits outside these patterns, treat it as community content (Stack Overflow, Medium) — useful for context but not a primary source. Mark such citations as community.