agentleFS
jramsahaijramsahai/obsidian-mcp-brain/vault-recall/SKILL.md

vault-recall

Skill0 starsChanged 2 months agoMITView on GitHub

Answer questions from the Obsidian vault: "what do I know about X", "what did I discuss with [person]", "when/why did we decide Y", "summarize my week", project or topic history. Read-only search, reading, and synthesis across notes, links, and syntheses. NOT for: capturing or changing anything (use the routing skills), daily standup (standup), or vault reorganization (nightly-consolidation).

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 2 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use second-brain for vault conventions and the shared tool surface.

    line 16
  2. Distinctive:

    Every tool here is read-only.

    line 20
  3. Very specific:

    Named entity first: If the question names a person, project, store, topic, or idea, read its note directly with obsidian__vault_read …

    line 36
  4. Distinctive:

    Never fill a gap from general knowledge without labelling it as outside the vault.

    line 56

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Vault Recall

Answer the user's questions from what the second brain already holds. Read-only: recall never mutates the vault. If the answer prompts a change ("actually, mark that done"), route it to the owning skill afterwards.

Use second-brain for vault conventions and the shared tool surface. An empty search result and a "not found" are authoritative answers, not errors — see second-brain -> Tools.

Tools

Every tool here is read-only. Never construct a file path: pass the note name as it appears in a wikilink and the server resolves it.

NeedCall
Read a noteobsidian__vault_read note="Wayfinder"
Read one sectionobsidian__vault_read note="Jane Doe" section="Conversation History"
Full-text searchobsidian__vault_search query="pricing"
Narrow a searchobsidian__vault_search query="pricing" type="knowledge"
Notes by type, folder, status, or change dateobsidian__vault_list type="project" status="Active"
Newest note in a dated folderobsidian__vault_list folder="Syntheses" latest=true
Backlinks — what links hereobsidian__vault_links direction="in" note="Wayfinder"
Outgoing linksobsidian__vault_links direction="out" note="Wayfinder"
Orientation before a broad questionobsidian__vault_status

Search Strategy (layered — stop as soon as the question is answered)

  1. Named entity first. If the question names a person, project, store, topic, or idea, read its note directly with obsidian__vault_read note="…". The note and its tables usually answer the question outright.
  2. Search. obsidian__vault_search on the key terms; try 2-3 phrasings before concluding absence. Narrow with type or folder when you know where the answer lives.
  3. Follow the graph one hop. obsidian__vault_links direction="in" and direction="out" on the notes found, then read their ## Related sections. This catches context that keyword search missed.
  4. Time questions use time sources. For a date range, list the daily and synthesis notes in range with obsidian__vault_list folder="Daily" changed_since="2026-07-01", then read them — plus completed tasks and dated table rows.

A search that returns nothing has answered the question. Do not re-run it a third way.

Recipes by Question Shape

  • "What do I know about TOPIC" -> obsidian__vault_search query="TOPIC" type="knowledge", then the topic's MOC note, then the notes it lists. Answer from the notes and name the ones used.
  • "What did PERSON and I discuss / what's pending with them" -> obsidian__vault_read note="First Last" section="Conversation History" and section="Pending Topics", then obsidian__vault_links direction="in" note="First Last" for project notes that mention them.
  • "Why/when did we decide X" -> the project note's ## Key Decisions table first, then obsidian__vault_search.
  • "Summarize my week/month" -> daily notes and syntheses in range, tasks completed in range, project ## Conversation Log rows in range.
  • "What am I waiting on / what's stale" -> that is standup's shape. Run standup rather than reinventing it.

Answer Style

  • Answer first, then sources: name the notes used so the user can jump in.
  • Quote the user's own recorded wording for decisions and journal content; do not paraphrase it into new claims.
  • "The vault says nothing about this" is a real answer — say which places were checked. Distinguish it from "found but thin."
  • Never fill a gap from general knowledge without labelling it as outside the vault.

Privacy

Journal and people-note content is sensitive: answer in the current direct channel only, include the minimum detail that answers the question, and never surface private details in group or public channels.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.