query
/query
Answer questions about the wiki by searching, reading, and synthesizing relevant scraps with `[[Title]]` citations. Use this when the user wants to query the wiki, ask what they have written about a topic, compare scraps, find related notes, or pull a synthesized overview from existing scraps.
Assessment and security
- Files in this package
- 1
- Scripts
- 0
- HTML comments
- 0
- URLs in SKILL.md
- 2
No signals detected. Repository changed on 27.10.2026, without references to retired models or APIs.
MIT, open: use and modify while retaining author attribution.
What it does and when to use it
Answer questions about the wiki by searching, reading, and synthesizing relevant scraps with `[[Title]]` citations. Use this when the user wants to query the wiki, ask what they have written about a topic, compare scraps, find related notes, or pull a synthesized overview from existing scraps.
Arguments
/query [question]
Original author description
Answer questions about the wiki by searching, reading, and synthesizing relevant scraps with `[[Title]]` citations. Use this when the user wants to query the wiki, ask what they have written about a topic, compare scraps, find related notes, or pull a synthesized overview from existing scraps.
How it works
- When to use
- Workflow
- Citation rules
- CLI used
- Further reading
SKILL.md preview
---
name: query
description: Answer questions about the wiki by searching, reading, and synthesizing relevant scraps with `[[Title]]` citations. Use this when the user wants to query the wiki, ask what they have written about a topic, compare scraps, find related notes, or pull a synthesized overview from existing scraps.
allowed-tools: Read, Bash, Glob, Grep
user-invocable: true
argument-hint: [question]
---
# Query
Answer a question against the wiki and return a citation-rich synthesis.
Implements Karpathy's *Query* primitive for Scraps: search the wiki, read relevant scraps, and synthesize an answer that references its sources by `[[Title]]`. Citations make the answer auditable and ready to be filed back as a new scrap if the user chooses.
## When to use
- "What do I have on X?" / "What did I write about Y?"
- "Compare X and Y across my notes"
- "Summarize what I know about Z"
- "Find scraps related to W"
## Workflow
1. **Understand the question**
- Identify the core topic and any constraints (time range, ctx folder, tag, etc.)
- If the question is ambiguous, ask one clarifying question before searching
2. **Search (broad → narrow)**
- `scraps search "<keyword>" --json` for each main keyword
- Try multiple phrasings if the first search returns few results
- For tag-driven questions: `scraps tag backlinks "<tag>" --json`
3. **Select candidates**
- From search results, pick the 5–15 most relevant scraps
- Prefer scraps that span the question (different ctx, different tags) over many near-duplicates
4. **Read selected scraps**
- `scraps get "<key>" --json` for each candidate, passing the hit's `key` back
- Use field projection to save context when appropriate, e.g. `--json body`, `--json headings`, or `--json code_blocks`
- If a link result includes `heading`, read just that section with `scraps get "<key>" --heading "<heading>" --json body`
- For graph-shaped questions, also use:
- `scraps links "<key>" --json` (outbound)
- `scraps backlinks "<key>" --json` (inbound)
5. **Synthesize with citations**
- Write the answer in plain Markdown
- Cite every claim that comes from a scrap as `[[Title]]` (or `[[Ctx/Title]]` when needed)
- Do not invent information beyond what the scraps and the user's question support
- Choose the output shape that fits the question:
- prose paragraph for narrative questions
- GFM table for comparisons
- bullet list for enumerations
- mermaid diagram for relationships
6. **Stop**
- Do not auto-offer to save the answer. If the user wants the synthesis saved as a new scrap, they invoke the `ingest` skill on the answer markdown.
## Citation rules
- Every non-trivial claim should trace to a `[[Title]]` citation.
- If a claim has no source in the wiki, mark it as outside the wiki ("not in the current scraps"); do not bluff.
- Cite with the exact `key` value returned by `scraps search --json` / `scraps get --json`, as `[[key]]`, so the citation resolves cleanly. The `title` is what readers see and is not a link target.
- A key already carries its context folder (`[[Ctx/Name]]`).
- `scraps links --json` may return a `heading`; use it for targeted reads, but cite the scrap as `[[Title]]` or `[[Ctx/Title]]` unless the answer specifically needs a heading reference.
## CLI used
- `scraps search <query> [--logic and|or] --json`
- `scraps get <key> [--heading <heading>] --json [fields]` (`<key>` is the scrap's path without `.md`, as results return it; `<name> --ctx <ctx>` is still accepted)
- `scraps links <key> --json` (returns `{ kind, key, title, ctx, heading }` refs)
- `scraps backlinks <key> --json`
- `scraps tag list --json`
- `scraps tag backlinks <tag> --json`
- `scraps todo [--status open|done|deferred|all] --json`
These commands are shown with `--json` because this skill consumes structured output — without it Scraps prints human-formatted text that is unstable to parse, so `--json` is part of the command, not an option. Use `scraps <cmd> --help` only as a fallback for a rare flag not listed here, not as a routine discovery step; for deeper syntax or spec questions, consult the `scraps-llm-wiki-schema` agent (it reads the official docs).
## Further reading
- Scraps docs: <https://boykush.github.io/scraps/>
- Karpathy's *Ingest / Query / Lint* framing: <https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f>
- Composition with other primitives, dialogue / catch-up patterns, and primitive boundaries: see the `scraps-llm-wiki-schema` agent.
Versions
| Repository | Score | Licence |
|---|---|---|
| This page · boykush/scraps | 90 | MIT |
| h1994st/rllvm | 100 | Apache-2.0 |
| garrytan/gbrain | 100 | MIT |
| garrytan/gbrain | 100 | MIT |
| proteanhq/protean | 100 | Apache-2.0 |
| oxy-hq/airlayer | 100 | Apache-2.0 |
| erayozenn/BigHead-2ndBrain | 100 | MIT |
| lambdalisue/cclens | 100 | MIT |
| oxy-hq/airlayer | 100 | Apache-2.0 |
Database snapshot: 13.11.2026 · package version 2b7b16123ab8 (SHA-256 digest of the entire package)
Downloads in the last 30 days: 0 · ranking
Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.
Similar skills
- 90wiki-query
Answers a question against an Obsidian-based LLM knowledge wiki. Performs index-first search, reads relevant pages, synthesizes an answer with citations, and optionally files the synthesis back as a new wiki page. Triggers when the user runs /wiki:query or asks a question that… - 100wiki-query
Ask the Obsidian LLM wiki (wiki/concept/) a question and get a synthesized answer with [[ ]] sources; good cross-page answers go back into the wiki. Use when asking the wiki (wiki に聞いて), or when past daily-research notes would answer faster than grep. - 100wiki-query
Pull from your accumulated business knowledge instead of starting from scratch. Triggers when the owner says "what do I know about X", "what's my thinking on Y", "pull up the wiki on Z", "what's in my wiki about [topic]", or runs /wiki-query. Reads the wiki index, identifies… - 100wiki-query
Query the wiki to answer questions. Searches wiki pages, synthesizes answers with citations, and optionally files valuable answers back as new wiki pages. Use when the user asks a question about the knowledge base. - 88wiki-query
Answer a question using the compiled wiki knowledge base, synthesising a response with [[wikilink]] citations. Always use this skill when the user says /wiki-query, 'what does my wiki say about', 'what do I know about', 'check my notes on', 'search my wiki for', or 'what's the… - 100wiki-query
Use when asking a question against a personal wiki built with wiki-init and wiki-ingest. Do not answer from general knowledge — always read the wiki pages first.