SuperAPI AI Directory

query

Writing & research

View repository

Updated 10-15 22:39

About this skill

Query an LLM Wiki

Answer from the maintained wiki, then capture only durable synthesis that makes future queries more useful.

Workflow

  1. Find the vault and read wiki-purpose.md; confirm that the question falls within its intended scope. Read wiki-schema.md for local conventions.
  2. Search with llm-wiki search "<question>". Use --bm25-only only when vector search is unavailable or inappropriate. Scan wiki/ for exact terms if needed.
  3. Read the matched pages and follow relevant [[wikilinks]] and ## Related sections. Use the graph as context rather than treating search ranking as the answer.
  4. Give a direct answer grounded in the pages. Cite pages as [[slug]], state uncertainty, distinguish evidence from inference, and surface conflicts or gaps instead of guessing.
  5. If the wiki lacks enough evidence, say so and suggest specific material to ingest. Do not fill gaps from unsaved external knowledge.

When knowledge should compound

Write back only a non-trivial, high-confidence synthesis: for example, a new connection among at least three pages, a resolved contradiction, or a useful comparison absent from all source pages. Do not create a page for a simple lookup or speculative conclusion.

For a durable synthesis, create a focused wiki page with complete frontmatter:

---
title: Synthesis Title
description: One-line summary
tags: [synthesis]
sources: [contributing-page-slugs]
source_type: query-synthesis
created: YYYY-MM-DD
updated: YYYY-MM-DD
---

Link it to the contributing pages, append a query entry to wiki-log.md, and run llm-wiki sync. Do not modify anything under sources/.