agentleFS
tboome33tboome33/obsidian-mcp-router/skills/wiki-tour/SKILL.md

wiki-tour

Skill1 starsChanged 7 months agoApache-2.0View on GitHub

Generate a guided, pedagogical reading tour through a vault — an ordered walkthrough that takes a newcomer from "what is this?" to "I get how it fits together", computed from the knowledge graph's link topology. Use when the user says "give me a tour of this vault", "where do I start", "walk me through my notes", "onboarding path", "/wiki-tour", "fais-moi visiter le vault", "parcours guidé", "par où commencer", or wants a learning path through a vault or one of its projects/topics.

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 3 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    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
  1. Distinctive:

    No arg → whole-vault tour.

    line 21
  2. Very specific:

    If it warns no-articles-in-scope / no-tour-steps, tell the user the scope is empty and offer alternatives.

    line 32
  3. Distinctive:

    Keep it 5-15 steps. Each step highlights 1-5 nodes (the skeleton already caps this).

    line 42
  4. Very specific:

    Don't invent step content — narrate from the nodes[].summary the tool returns.

    line 61
  5. Distinctive:

    Don't reorder the deterministic skeleton arbitrarily — its order (overview → topics by importance) is the pedagogical spine.

    line 62
  6. Very specific:

    Don't overwrite the whole-graph tour[] from a *scoped* tour.

    line 63
  7. Very specific:

    Don't exceed ~15 steps — a tour is a curated path, not a table of contents (that's catalog.md).

    line 64

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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

wiki-tour

Build a guided tour through a vault: an ordered sequence of steps that teaches the vault's structure in a sensible reading order, derived from the knowledge graph's link topology (fan-in/backlinks, entry points, catalog.md sections). Reference: [[understand-anything-roadmap]] item #3.

Same split as /wiki-graph: the topology + step ordering is deterministic (the build_wiki_tour tool); you (Claude) write the pedagogical narrative — the why of each step.

Pre-condition

The knowledge graph must exist: wiki-meta/graph/knowledge-graph.json. If it doesn't, run /wiki-graph first (the tool will tell you). Richer tours come from a richer graph — suggest /wiki-refresh-digests + /wiki-graph if the vault has few digests.

Steps

1. Resolve scope

  • No arg → whole-vault tour.
  • scope arg → one section/topic (by catalog.md heading) or a path substring. For a multi-project vault, tour one project at a time (e.g. /wiki-tour Dedibox). If unsure which scopes exist, read wiki-meta/catalog.md section headings and offer a menu.

2. Compute the deterministic skeleton

mcp__obsidian-router__build_wiki_tour({ vault, scope?, maxSteps? })

Returns: ordered steps[] — each with order, a default title, and nodes[] ({id, name, summary}) — plus entryPoints[] and totalArticles. The skeleton is: an overview step (the entry points — highest fan-in, boosted for index/MOC names) then one step per catalog.md section (its top articles by backlink count), then a trailing "other notable pages" step for unindexed hubs.

If it warns no-articles-in-scope / no-tour-steps, tell the user the scope is empty and offer alternatives.

3. Write the pedagogical narrative (LLM — your job)

For each skeleton step, write:

  • A clear step title (refine the default if needed).
  • A 2-4 sentence description: what this step covers, why it matters, and how it connects to the previous step — building a coherent "what is this → how it works" arc. Ground it in the nodes[].summary provided; don't invent.
  • Follow the vault's language convention — the languages value in its conventions file (one language: everything in it; several: one section per language, in order). A vault that still carries the retired bilingual convention (not yet migrated) is governed by that section's own text — French then English — until its owner migrates it; the migration value is theirs to choose and may differ.
  • Optionally a languageLesson-style aside for domain concepts, if useful.

Keep it 5-15 steps. Each step highlights 1-5 nodes (the skeleton already caps this).

4. Output — BOTH

  1. Standalone markdown tour page (usable in Obsidian today): write to wiki-meta/tours/<scope-or-vault>-tour.md with frontmatter (type: tour, scope, generated_at) + an ordered list of steps, each linking its nodes as [[wikilinks]] (so the reader clicks through). This is the immediately-useful artifact.
  2. Graph tour[] field (for the future dashboard / #2b viewer): read wiki-meta/graph/knowledge-graph.json, set its tour to the narrated steps ([{order, title, description, nodeIds}]), write it back. Keep the rest of the graph intact. (Only when scope = whole-vault — a scoped tour shouldn't overwrite the full-graph tour; for a scoped tour, write only the markdown.)

5. Confirm

Report: scope, step count, output path, and a one-line preview of the arc (step titles).

When NOT to use

  • A single page's neighbours → get_wiki_context_pack.
  • A portable text dump → wiki-export.
  • The graph doesn't exist yet → /wiki-graph first.

Anti-patterns

  • Don't invent step content — narrate from the nodes[].summary the tool returns.
  • Don't reorder the deterministic skeleton arbitrarily — its order (overview → topics by importance) is the pedagogical spine; refine titles/descriptions, not the sequence, unless you have a clear pedagogical reason (then say why).
  • Don't overwrite the whole-graph tour[] from a scoped tour.
  • Don't exceed ~15 steps — a tour is a curated path, not a table of contents (that's catalog.md).

Quirks

  • Deterministic skeleton: same graph ⇒ same step order (only your narrative varies).
  • scope matches a catalog.md section name, a layer id, or a path substring — first match wins.
  • Entry points are boosted for names like index/overview/MOC/sommaire; otherwise highest backlink count wins.

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.