llmwiki-query

name	llmwiki-query
description	Query the LLM-maintained knowledge wiki and synthesize a focused answer. Searches _wiki/ pages first, falls back to raw vault notes, and files new insights back into the wiki automatically — so every query compounds the knowledge base.
version	1.0.0
metadata	{"openclaw":{"emoji":"🔍","homepage":"https://github.com/wangsw/llm-wiki-skills","always":false}}

/wiki-query — Query the LLM Wiki

You are an expert knowledge synthesizer. Your job is to answer questions by leveraging the accumulated _wiki/ knowledge layer — not re-deriving everything from scratch — and to file new insights back so the wiki compounds over time.

Trigger

User invokes /wiki-query followed by a question or topic:

/wiki-query what is the difference between X and Y
/wiki-query how does the authentication flow work
/wiki-query summarize everything I know about machine learning

Workflow

1. Search the wiki

   • Grep _wiki/ for keywords from the question
   • Try synonyms and related terms if first search is sparse

2. Read relevant pages

   • Read the most relevant pages found
   • Follow [[wikilinks]] to related pages for additional context

3. Fall back to the vault if needed

   • If wiki coverage is thin, Grep the full vault for raw notes
   • Raw notes are secondary — the wiki is the primary source

4. Synthesize an answer

   • Combine findings into a clear, direct response
   • Cite sources as _wiki/page-name.md or vault file paths

5. File new insights back

   • If you derived something new not already in the wiki, update or create the relevant page
   • This is how the wiki compounds over time

6. Report gaps

   • State explicitly what the wiki is missing
   • Suggest which sources to run /wiki-ingest on

Output Format

<direct answer>

Sources: _wiki/page-name.md, _wiki/other-page.md

Wiki updated: yes — added [what] to [page] / no
Gaps: [missing coverage, if any]

Use Cases

• Answer questions using accumulated personal knowledge
• Summarize everything known about a topic
• Find connections between concepts across the vault
• Identify gaps in the knowledge base
• Retrieve past decisions or research without re-reading raw notes

References

• Wiki Page Format
• Frontmatter Schema