vault-query
Skill2 starsChanged 7 months agoMITView on GitHub
Answer a research question using the wiki. Use when the user asks a question of the form "what do my sources say about X", "where do they disagree", "summarize X and cite sources", or any question the wiki might already have synthesized. Reads hot.md, then index.md, then 5–7 relevant pages (synthesized pages before raw sources). Offers to file the answer as an Outputs/ page so future queries compound. Do NOT use for ingesting new sources (use vault-ingest) or reviewing queue (use vault-review).
4 other projects use a version of this.
Adopted by
Most recently taken up by
- bohanyt/hermes-custom-skillstheir own version, first seen Oct 25, 2026
- Lucioway/way-stacktheir own version, first seen Aug 31, 2026
- quietdaemon/starter-skillstheir own version, first seen Jun 6, 2026
- Howie126313/obsidian-vault-kittheir own version, first seen Apr 13, 2026
Other versions4 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 4
- Howie126313/obsidian-vault-kit★ 3same name
- bohanyt/hermes-custom-skills★ 1same name
- quietdaemon/starter-skills★ 1same name
- Lucioway/way-stack★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository11 other files it gives its agents.
- AGENTS.mdAGENTS.md
- agent-browserSkill
- defuddleSkill
- json-canvasSkill
- obsidian-basesSkill
- obsidian-cliSkill
- obsidian-markdownSkill
- vault-ingestSkill
- vault-initSkill
- vault-maintainSkill
- vault-reviewSkill
The file
We split this skill into the rules it gives an agent
8 rules. 0 are common, 4 are distinctive, and 4 are very specific to this project.
0
Common
Most of the 4 edited versions we read keep these.
4
Distinctive
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 30
Read hot.md (session cache) first
- Very specific: line 31
Read index.md (navigation hub)
- Distinctive: line 37
Prefer synthesized pages (Concepts/, Entities/, Outputs/) before raw Sources/.
- Distinctive: line 49
When composing the response:
- Distinctive: line 63
If the answer is substantive and potentially reusable, ask: *"Should I file this as an Outputs/ page so future queries can reuse it?"*
- Very specific: line 65
If yes, create Outputs/<Short Title>.md using _templates/Output.md:
- Very specific: line 91
If an Outputs/ page matures (referenced repeatedly, stabilized), consider promoting its durable insights into Concepts/ or Entities/ via …
- Distinctive: line 96
Don't synthesize past the evidence: If two sources say X and one says Y, don't present X as "most agree" — present the disagreement.
“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Vault Query Skill
Answer a research question against the existing wiki and — critically — file the answer back so future queries compound instead of re-deriving.
When to Use
Trigger on:
- "What do my sources say about
<topic>?" - "Where do the sources disagree on
<topic>?" - "Summarize
<Concept>and cite every source." - "Is there an answer in the wiki for
<question>?" - Any research question that isn't obviously an ingest or maintenance request
When NOT to Use
- The user is adding a new source (use
vault-ingest) - The user is reviewing the queue (use
vault-review) - The user wants a general answer not tied to the wiki (respond normally without this skill)
Workflow
Step 1 — Restore context
- Read
hot.md(session cache) first - Read
index.md(navigation hub)
Token budget: ~1,500 tokens combined. If the vault is past 50 pages, index.md may be MOC-style with topic indexes — follow those to the right sub-index.
Step 2 — Route to the right pages
Prefer synthesized pages (Concepts/, Entities/, Outputs/) before raw Sources/. Reading a Concept page that synthesizes 5 sources is more efficient than reading all 5 sources.
Limit: 5–7 pages per query round. If you need more, either narrow the question with the user or file partial answers as Outputs/ pages and iterate.
Tools available for finding relevant pages:
rg <term>for keyword search across the vaultbin/stats.shfor a high-level view if you're not sure where to look- Aliases in frontmatter (a page titled "Machine Learning" with
aliases: [ML]resolves both)
Step 3 — Synthesize the answer
When composing the response:
- Cite every substantive claim with a wikilink to the source:
... per [[Sources/<Title>]] - Flag disagreements explicitly with
> [!contradiction]if sources conflict - Distinguish extracted vs. inferred claims — if you're synthesizing beyond what sources state, mark with
> [!info] Inferred - Note gaps — if the wiki can't answer part of the question, say so (and suggest an ingest)
- Respect confidence — don't promote a
confidence: lowclaim to unqualified fact
Step 4 — Answer in chat first
Give the user the answer in the conversation. Don't file as an Output before they've seen it — they may want to refine.
Step 5 — Offer to file the answer
If the answer is substantive and potentially reusable, ask: "Should I file this as an Outputs/ page so future queries can reuse it?"
If yes, create Outputs/<Short Title>.md using _templates/Output.md:
---
query: "<the original question>"
status: current
created: <YYYY-MM-DD>
---Sections:
- Question — the original question
- Findings — the synthesized answer with citations
- Methodology — which pages were consulted, how you decided on the answer
Step 6 — Cross-link
Link the new Outputs/ page from the Concepts/ and Sources/ pages it draws on. Those pages now point to the filed answer — future readers find the synthesis without re-deriving.
Step 7 — Update log and hot cache
log.md: append## [date] query | "<question>"— note the output page createdhot.md: add to Recent Changes if notable; add to Active Threads if this opened a new research direction
Crystallization Trigger
If an Outputs/ page matures (referenced repeatedly, stabilized), consider promoting its durable insights into Concepts/ or Entities/ via the vault-maintain skill's crystallization workflow.
Anti-Patterns
- Don't skip the index. Reading 20 sources unfiltered is a sign you didn't route through synthesized pages first.
- Don't synthesize past the evidence. If two sources say X and one says Y, don't present X as "most agree" — present the disagreement.
- Don't file every chat answer. Some questions are one-offs. File when the answer is reusable or the synthesis took work.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

