po4yka/llm-wiki-skills / skills/llm-wiki-setup
llm-wiki-setup
VerifiedInstall, configure, and connect a chosen LLM-Wiki workflow. Use for setting up a raw/wiki/schema vault, Agent Skills, coding-agent integration, Obsidian/local-first workflows, hooks, templates, or git review; route a guided first-demo walkthrough to llm-wiki-zero-to-working-wiki.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 4.2 KB |
| references/templates/vault/AGENTS.md | markdown | 1.4 KB |
| references/templates/vault/CLAUDE.md | markdown | 1.2 KB |
| references/templates/wiki/index.md | markdown | 483 B |
| references/templates/wiki/log.md | markdown | 400 B |
SKILL.md
View on GitHub---
name: llm-wiki-setup
description: Install, configure, and connect a chosen LLM-Wiki workflow. Use for setting up a raw/wiki/schema vault, Agent Skills, coding-agent integration, Obsidian/local-first workflows, hooks, templates, or git review; route a guided first-demo walkthrough to llm-wiki-zero-to-working-wiki.
license: MIT
compatibility: Designed for Agent Skills-compatible coding agents. Requires filesystem access for local setup; may require internet access for installing external tools.
metadata:
author: po4yka
version: "1.0.0"
install_scope: self-contained
---
# LLM-Wiki Setup
## Goal
Turn a chosen LLM-Wiki recommendation into a working setup with safe defaults.
## When to use
- The user has already picked a target vault/architecture and wants it turned into a working setup (directories, templates, skills installed).
- Configuring `AGENTS.md`/`CLAUDE.md`, hooks, or git/PR review policy for an LLM-Wiki workflow.
- Connecting a coding agent (Claude Code, Codex, Cursor, OpenCode) to Agent Skills so `wiki-ingest`/`wiki-query`/`wiki-lint` work.
- Migrating an existing Obsidian vault or repository docs tree into the raw/wiki/schema model.
- Not for a guided first-demo walkthrough with no prior architecture decision — route that to `llm-wiki-zero-to-working-wiki` instead.
## Inputs
- Chosen solution or target architecture.
- Project/vault path.
- Target agent(s): Claude Code, Codex, Cursor, OpenCode, etc.
- Privacy/local-first constraints.
- Existing docs or empty start.
## Procedure
### 1. Confirm the setup target
Identify one target:
- local-first Markdown vault;
- existing Obsidian vault;
- repository documentation wiki;
- qmd/hybrid-search upgrade;
- team PR-based workflow;
- custom product/CLI/plugin.
If the target is an external tool, browse official installation docs before giving commands.
### 2. Preflight
Check or ask for:
- git initialized;
- backup or branch available;
- package managers available;
- agent skill installation target;
- whether writes should be direct, dry-run, or PR-based;
- sensitive material that must not leave the machine.
### 3. Create the vault shape
Default structure:
```text
raw/sources/
raw/assets/
inbox/
wiki/index.md
wiki/log.md
wiki/sources/
wiki/entities/
wiki/concepts/
wiki/comparisons/
wiki/synthesis/
wiki/queries/
_agent/reports/
_meta/schemas/
```
Start `wiki/index.md` and `wiki/log.md` from the bundled `references/templates/wiki/` starters; prefer the target repository's own vault templates when they already exist.
### 4. Install skills
For Agent Skills-compatible agents, use commands like:
```bash
npx skills add po4yka/llm-wiki-skills --skill '*' -a claude-code
npx skills add po4yka/llm-wiki-skills --skill wiki-ingest --skill wiki-query -a codex
```
Use project scope when the skills should be committed with a repo; use global scope when the user wants them across projects.
### 5. Configure instructions
Create or update `AGENTS.md` and `CLAUDE.md`, starting from `references/templates/vault/AGENTS.md` and `references/templates/vault/CLAUDE.md`, keeping:
- raw/wiki/schema model;
- protected human sections;
- prompt-injection defense;
- git/dry-run policy;
- index/log update rules;
- allowed write scope.
### 6. Add safety checks
Recommend, and implement only with approval:
- git branch before migration;
- frontmatter validation;
- dry-run mode for bulk changes;
- pre-write checks for `raw/` and protected sections;
- `wiki-lint` report after setup.
### 7. Verify
Run a tiny end-to-end test:
1. Put one sample source in `raw/sources/` or `inbox/`.
2. Run `wiki-triage` or `wiki-ingest`.
3. Run `wiki-query` against the result.
4. Run `wiki-lint` and inspect the report.
## Output
Deliver the summary in chat; the persisted artifacts are the created vault files and installed skills.
```markdown
## Setup summary
## Files created or changed
## Install commands used
## Safety settings
## Verification result
## Next action
```
## Safety gates
- Do not run remote install scripts without explaining what they do.
- Do not ingest sensitive data into cloud tools without explicit approval.
- Do not enable bulk direct writes before git and backup are confirmed.
- Do not combine setup with large migration unless the user requested it.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |