To Wiki logo

To Wiki

Community
elianiva
to-wiki

Take knowledge from the current conversation and compound it into a living wiki in the personal Obsidian vault. The agent scouts existing pages via QMD first, proposes what to write, quizzes the user, then upserts.

Overview

Publisherelianiva
Repositorydotfiles
Skill nameto-wiki
Stars
204
Forks
9
Bundled files
Instructions only
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by elianiva on GitHub. Read the source before you install it.

Installation

Install the To Wiki AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/elianiva/dotfiles.git /tmp/dotfiles
mkdir -p .claude/skills
cp -r /tmp/dotfiles/agents/skills/to-wiki .claude/skills/to-wiki
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable To Wiki in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads To Wiki on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so To Wiki is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

To Wiki

Turn what you and the user have been discussing into a living wiki — a persistent, interlinked collection of markdown pages that grows richer with every conversation. You do the reading, writing, and cross-referencing; the user curates what's worth keeping.

Everything lives in the personal vault at ~/Development/personal/notes/. The only variable is which directory within it:

DirectoryWhen
Vault/<topic>/General knowledge — concepts, techniques, definitions, reference material.
Projects/<project>/Knowledge about a specific project — architecture decisions, domain concepts, code patterns, findings.

The skill picks the directory based on context. If the conversation is about code in a known project, it goes in Projects/<project>/. If it's general, it goes in Vault/<topic>/. When ambiguous, it asks.

Conventions

Frontmatter

yaml
---
tags:
  - <kebab-case-tag>
  - <topic-area>
created_at: <YYYY-MM-DD>
aliases: []
id: <Title as in filename>
---

Rules:

  • id matches the filename (minus .md). This is how [[Wiki-Links]] resolve.
  • created_at is set once on creation. Add updated_at on subsequent edits.
  • Tags are kebab-case. Reuse existing tags from the vault where possible.
  • #public is never added automatically. Only when the user explicitly confirms a page is for the digital garden. Everything is private by default.

Writing style

  • Structured markdown with ### subsections.
  • Bullet lists for takeaways or enumerations.
  • Code blocks with language tags for technical content.
  • Every page links to at least one other page. A page with zero links is an orphan — orphan pages get lost. Link from new pages to existing ones, and offer to add backlinks from existing pages to new ones.

Inside Projects/

Each project gets its own directory under Projects/:

notes/
└── Projects/
    └── <project-name>/
        ├── index.md          # Catalog — auto-updated on every upsert
        └── <Topic>.md        # Knowledge pages

The project name matches the directory name you use for the project (e.g. remix-app, lutra). The agent figures out which project you're working on from the current working directory, the files in context, or your conversation.

Process

1. Scout the terrain

Search the existing wiki before writing anything. Do this autonomously — don't ask the user first.

Use QMD to search broadly first — it covers the whole vault:

bash
qmd query "<topic>" -c vault

Then narrow down based on what you find:

  • If the knowledge relates to a project, also search within Projects/<project>/ using QMD or by reading the directory.
  • If the topic has an obvious home (e.g. Vault/Software Engineering/ for a dev concept), scan that directory too.
  • Read any relevant pages in full with qmd get or the read tool.

Also assess the conversation context to determine which directory the content belongs in:

  • Is the conversation about code in a specific project (visible from cwd, open files, or what you're discussing)? → Projects/<project>/
  • Is it a general concept, technique, or reference? → Vault/<topic>/
  • If Vault/<topic>/ has an existing subdirectory that matches the content (e.g. Islam/, Motorsports/, Software Engineering/), use it. If none fits and there's no clear project, place it in Vault/ directly.

Completion criterion: You know what exists (or that nothing does), have identified the right directory for the new content, and can articulate whether this is a creation, revision, or cross-reference addition.

2. Formulate

Based on scouting, determine the shape of the work:

  • Create — a new concept not yet in the wiki. Choose a title distinct from existing pages.
  • Update — an existing page needs new sections, revised content, or corrected facts. Changing updated_at is required.
  • Cross-reference — existing pages should link to each other, or a new page needs links to existing ones. Every page needs at least one link.

Also determine:

  • Title — concise, matches the concept name the user used.
  • TargetProjects/<project>/ or Vault/<topic>/. For project content, also determine the project name (from cwd, git remote, or files in context).
  • Key information — distil what matters from the conversation. The wiki is compiled knowledge, not a transcript. Keep the essential, drop the ephemeral.

Completion criterion: You have a concrete proposal: what page(s) to touch, where they go, and what changes to make.

3. Present to the user

Show your proposal concisely:

→ **<Page Title>** in Projects/<project>/
  New page: <1-2 sentence summary>
  Links to: [[Existing Page]], [[Another Page]]

Or for general content:

→ **<Page Title>** in Vault/<topic>/
  New page: <1-2 sentence summary>
  Links to: [[Existing Page]], [[Another Page]]

Ask only what the user needs to decide:

  • "The title I'm thinking is X — does that work?"
  • "This is related to Y which already has a page — merge or keep separate?"
  • "I found these existing cross-references — should I add backlinks?"

Keep it tight. One round of confirmation resolves most cases. If the proposal is straightforward (new page, no ambiguity), a single "This looks right — write it" from the user is enough.

Completion criterion: The user has confirmed the plan. If there was ambiguity, one exchange resolves it.

4. Upsert

Write or update the pages with the agreed conventions.

New page — use the write tool to create the file with full frontmatter, content, and wiki-links.

Existing page — read it first with read, then use edit to update specific sections, or rewrite the whole file if changes are extensive. Always update updated_at in frontmatter.

Cross-references are a two-way street. After writing, check if pages you linked to should link back. Offer this to the user rather than doing it silently.

Update the index. If index.md exists, add or update the entry for every page you touched. If it doesn't exist and this is the first real page, create it:

markdown
# Wiki Index

## <Category>

- [[<Page Title>]] — <one-line summary>

Append to the appropriate category section. Create a new category if none fits.

Never touch:

  • Daily/ or Inbox/ — they are private and ephemeral by design.
  • Articles/ or People/ — they are for source summaries and profiles, not wiki knowledge. Only write there if the user explicitly asks.
  • Templates/, Images/, Music/ — each has a specialised purpose. The wiki lives in Vault/ and Projects/.

Never overwrite without confirmation. If a page already exists and you're proposing significant changes, show the diff to the user before applying.

Completion criterion: All approved pages are written or updated, with proper frontmatter, cross-references, and an updated index. The wiki is internally consistent — no orphan pages created, no stale index entries left.

5. Confirm

Summarise what was done:

Done. Created [[<Page>]] in project wiki, linked from [[<Related Page>]]. Index updated.

If anything is noteworthy — new tags created, a subdirectory was created for a dense topic, an existing page was significantly revised — mention it briefly.

Completion criterion: The user knows what happened and can verify the result.

What makes this different

  • Scout before write. Never assume the wiki is empty. Never overwrite knowledge that already exists.
  • Compile, don't transcribe. The wiki is distilled knowledge, not a chat log. If the user said something in passing that isn't worth keeping, leave it out.
  • Link or be orphaned. Every page connects to at least one other. The wiki's value is in the cross-references.
  • Index as table of contents. Without the index, the agent has to grep for everything. Maintain it on every upsert.

Frequently asked questions

What does the To Wiki AI skill do?

Take knowledge from the current conversation and compound it into a living wiki in the personal Obsidian vault. The agent scouts existing pages via QMD first, proposes what to write, quizzes the user, then upserts.

Why use To Wiki on TypingMind?

Because you install it once and use it with any model. To Wiki is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install To Wiki in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/elianiva/dotfiles/tree/master/agents/skills/to-wiki. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use To Wiki?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with To Wiki?

As many as you like. As long as a model supports skills, you can use To Wiki with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the To Wiki AI skill free?

It is published on GitHub by elianiva. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇