Imported from VectorPeak/LLM-Wiki (
AGENTS.md). Install upstream withnpx skills add VectorPeak/LLM-Wiki. Copyright stays with the author.
Obsidian Second Brain - Codex CLI Operating Manual
Local vault overrides: also follow
_CLAUDE.mdfor folder conventions and Markdown authoring rules. In particular,._trash&cache/is reserved for temporary stitching/cache files and soft trash, and Mermaid diagrams in new Markdown should default to theforesttheme with compact, whole-graph layouts.
This vault runs the obsidian-second-brain skill. The skill ships a set of commands: each one is a multi-step instruction file that you (the Codex agent) should follow when the user's request matches its trigger phrase.
How to operate
- Read
_CLAUDE.mdin the vault root, if it exists, to learn the user's vault conventions (folder map, daily note format, naming). - When the user's request matches a trigger in the tables below, read the
matching file under
.codex/commands/<name>.mdand follow its instructions step by step. - Treat the AI-first vault rule (
.codex/references/ai-first-rules.md) as non-negotiable for every note you write:## For future Claudepreamble, rich frontmatter (type,date,tags,ai-first: true),[[wikilinks]]for every person/project/concept, recency markers per external claim, sources verbatim, confidence levels where applicable. - Do not invent commands. If no command matches, ask the user what they want or fall back to plain natural-language help.
Command routing tables (grouped by category)
Wiki layer - simplified knowledge operations
| Command | What it does | Read this file |
|---|---|---|
/ingest |
LLM_wiki intake protocol - preserve raw sources, extract learning units, and update durable wiki knowledge | 02.wiki/projects/LLM_wiki Wiki Layer Protocol.md, then .codex/commands/ingest.md |
/query |
LLM_wiki retrieval protocol - search, explain, synthesize, and answer from vault evidence | 02.wiki/projects/LLM_wiki Wiki Layer Protocol.md, then .codex/commands/query.md |
/lint |
LLM_wiki maintenance protocol - audit links, duplicates, stale claims, raw backlog, DailyNotes drift, Mermaid defaults, and Mentor consistency | 02.wiki/projects/LLM_wiki Wiki Layer Protocol.md, then .codex/commands/lint.md |
Vault - daily writing, capture, find
| Command | What it does | Read this file |
|---|---|---|
/obsidian-board |
Show or update a kanban board - flags overdue items, updates from conversation | .codex/commands/obsidian-board.md |
/obsidian-capture |
Quick idea capture - zero friction, saves to the maintained capture area and mentions in daily note | .codex/commands/obsidian-capture.md |
/obsidian-daily |
Create or update today's daily note - pulls calendar events, overdue tasks, and conversation context | .codex/commands/obsidian-daily.md |
/obsidian-find |
Smart vault search - returns results with context, not just filenames | .codex/commands/obsidian-find.md |
/obsidian-log |
Log this work or dev session to the vault - infers project from context | .codex/commands/obsidian-log.md |
/obsidian-person |
Create or update a person note from conversation context | .codex/commands/obsidian-person.md |
/obsidian-project |
Create or update a project note - adds to board and daily note automatically | .codex/commands/obsidian-project.md |
/obsidian-projects |
Live project status from git + local docs - infers all context from vault notes, no config required | .codex/commands/obsidian-projects.md |
/obsidian-recap |
Summarize a time period from the vault - today, week, or month | .codex/commands/obsidian-recap.md |
/obsidian-recurring |
Track a recurring obligation (payment, filing, ops) with a cadence and a computed next-due date | .codex/commands/obsidian-recurring.md |
/obsidian-save |
Save everything worth keeping from this conversation to the vault | .codex/commands/obsidian-save.md |
/obsidian-task |
Add a task to the right kanban board with inferred priority and due date | .codex/commands/obsidian-task.md |
/obsidian-world |
Load your identity, values, priorities, and current state in one shot - with progressive context levels to avoid burning tokens | .codex/commands/obsidian-world.md |
Thinking - synthesis, decisions, learning, reviews
| Command | What it does | Read this file |
|---|---|---|
/idea-discovery |
Surface 3-5 next-direction candidates by reading ungraduated ideas, open project questions, and orphan research notes - what is worth working on next | .codex/commands/idea-discovery.md |
/obsidian-adr |
Generate a decision record when the vault structure changes - the vault knows why it knows what it does | .codex/commands/obsidian-adr.md |
/obsidian-challenge |
Red-team your current idea against your own vault history - finds contradictions, past failures, and flawed assumptions | .codex/commands/obsidian-challenge.md |
/obsidian-connect |
Bridge two unrelated domains using your vault's link graph - forces creative friction to spark new ideas | .codex/commands/obsidian-connect.md |
/obsidian-decide |
Extract decisions from this conversation and log them to the right project notes | .codex/commands/obsidian-decide.md |
/obsidian-emerge |
Surface unnamed patterns from your recent notes - recurring themes, hidden connections, and conclusions you haven't explicitly stated | .codex/commands/obsidian-emerge.md |
/obsidian-graduate |
Promote an idea fragment into a full project spec with tasks, board entries, and structure | .codex/commands/obsidian-graduate.md |
/obsidian-learn |
Review vault learnings, prune stale ones, surface active patterns - the vault's lessons compound or expire | .codex/commands/obsidian-learn.md |
/obsidian-panel |
Convene a panel of distinct perspectives on a decision - one independent verdict per lens, then a synthesis. A multi-persona complement to /obsidian-challenge | .codex/commands/obsidian-panel.md |
/obsidian-reconcile |
Find and resolve contradictions in the vault - the vault maintains its own truth | .codex/commands/obsidian-reconcile.md |
/obsidian-review |
Generate a structured weekly or monthly review note from vault history | .codex/commands/obsidian-review.md |
/obsidian-synthesize |
Automatic synthesis - scans the vault for unnamed patterns and writes synthesis pages without being asked | .codex/commands/obsidian-synthesize.md |
/vault-deep-synthesis |
Deep cross-reference of everything the vault knows about one topic - agreements, contradictions, stale claims, and coverage gaps. Pure vault, no network | .codex/commands/vault-deep-synthesis.md |
Research - bring external sources into the vault
| Command | What it does | Read this file |
|---|---|---|
/notebooklm |
Vault-first source-grounded research via Gemini File Search. One command, no browser. The grounded parallel to /research-deep (which is open-web via Perplexity). | .codex/commands/notebooklm.md |
/obsidian-ingest |
Ingest a source into the vault - the vault rewrites itself around new knowledge. Every ingest updates entities, rewrites stale claims, synthesizes new concepts, and resolves contradictions. | .codex/commands/obsidian-ingest.md |
/podcast |
Extract metadata, transcript, and summary from a podcast episode, saved as an AI-first note in the vault | .codex/commands/podcast.md |
/research |
Web research with citations - Perplexity Sonar when an API key is set, free key-less sources (Wikipedia, HackerNews, arXiv, Reddit, and more) otherwise. Deep dossier with summary, facts, timeline, players, contrarian views, open questions | .codex/commands/research.md |
/research-deep |
Vault-first deep research - scans the vault, fills gaps (Perplexity + Grok when keyed, free key-less sources otherwise), synthesizes a delta, then propagates updates across people/projects/ideas via /obsidian-save | .codex/commands/research-deep.md |
/x-pulse |
Scan X for what's trending in a topic - themes, voices, hooks, and post ideas powered by Grok + Live Search | .codex/commands/x-pulse.md |
/x-read |
Deep-read an X (Twitter) post via Grok + Live Search - verbatim post, thread, TL;DR, claims, reply sentiment, voices to watch | .codex/commands/x-read.md |
/youtube |
Extract transcript, metadata, and top comments from a YouTube video - summarized via Grok and saved to vault | .codex/commands/youtube.md |
Meta - vault setup, health, structure
| Command | What it does | Read this file |
|---|---|---|
/create-command |
Create a new obsidian-second-brain command via interview - zero markdown editing required | .codex/commands/create-command.md |
/obsidian-architect |
Scan a codebase and write a maintained set of architecture notes into the vault - overview, per-module notes, key decisions. Re-run to refresh without clobbering your edits | .codex/commands/obsidian-architect.md |
/obsidian-export |
Export a clean structured snapshot of the vault that any agent or tool can consume - flat JSON or markdown index | .codex/commands/obsidian-export.md |
/obsidian-health |
Run a vault health check - grouped by severity, detects contradictions, concept gaps, stale claims, and structural issues | .codex/commands/obsidian-health.md |
/obsidian-init |
Scan your vault and generate a _CLAUDE.md operating manual, index.md catalog, and .codex/references/log.md pointer |
.codex/commands/obsidian-init.md |
/obsidian-visualize |
Generate a visual canvas map of your vault - see the shape of your second brain and how knowledge connects | .codex/commands/obsidian-visualize.md |
Trigger phrases
When the user says any of the phrases below (or a close paraphrase), follow the matching command file.
English (en)
Wiki layer - simplified knowledge operations
/ingest- "ingest", "import to wiki", "add to wiki", "absorb into wiki", "导入", "吸收", "入库", "收进知识库", "剪藏入库"/query- "query wiki", "ask wiki", "search wiki", "explain from wiki", "what does my vault say", "查询", "问 wiki", "查一下知识库", "从我的笔记里回答", "解释这个概念"/lint- "lint wiki", "check wiki", "audit wiki", "vault lint", "knowledge base health", "体检", "检查 wiki", "检查知识库", "清理知识库"
Vault - daily writing, capture, find
/obsidian-board- "show board", "kanban", "what is on my board", "update board"/obsidian-capture- "capture this idea", "save this idea", "quick note", "drop a thought"/obsidian-daily- "todays note", "create todays daily", "open daily", "today daily note"/obsidian-find- "find in vault", "search my notes", "where is", "what did I write about"/obsidian-log- "log this work", "log this session", "log this dev session", "obsidian log"/obsidian-person- "save this person", "add person", "new contact note", "create person note"/obsidian-project- "new project", "create project note", "project setup", "start a project"/obsidian-projects- "projects overview", "project status", "what am I working on", "show projects"/obsidian-recap- "recap today", "recap the week", "summarize the week", "month recap"/obsidian-recurring- "recurring task", "monthly obligation", "remind me every month", "recurring payment", "track a recurring"/obsidian-save- "save this", "save the conversation", "save to vault", "obsidian save"/obsidian-task- "add task", "new todo", "track this", "remind me"/obsidian-world- "load context", "what is going on", "where am I", "load my world"
Thinking - synthesis, decisions, learning, reviews
/idea-discovery- "what should I work on next", "idea discovery", "surface next directions", "what's worth pursuing"/obsidian-adr- "log this decision", "ADR", "record decision", "decision record"/obsidian-challenge- "challenge this", "grill me on this", "red team my idea", "stress test this"/obsidian-connect- "connect domains", "cross-pollinate", "bridge ideas", "find an unexpected link"/obsidian-decide- "extract decisions", "log decisions", "what did we decide"/obsidian-emerge- "find patterns", "what is emerging", "surface themes", "unnamed patterns"/obsidian-graduate- "promote idea", "graduate this to project", "make a project from this", "elevate idea"/obsidian-learn- "review learnings", "what have I learned", "show lessons", "prune learnings"/obsidian-panel- "convene a panel", "advisor panel", "get multiple perspectives on", "panel review", "what would the experts say about"/obsidian-reconcile- "find contradictions", "reconcile vault", "fix conflicts", "vault contradictions"/obsidian-review- "weekly review", "monthly review", "review my week", "review my month"/obsidian-synthesize- "synthesize", "auto-synthesis", "make synthesis notes", "find unnamed patterns"/vault-deep-synthesis- "synthesize what I know about", "deep synthesis on", "cross-reference my notes on", "what does my vault say about"
Research - bring external sources into the vault
/notebooklm- "notebooklm", "research grounded", "ground research in vault", "ask my notebook", "source-grounded research"/obsidian-ingest- "ingest this source", "add this article", "import this", "absorb this"/podcast- "summarize this podcast", "podcast episode summary", "extract podcast", "what's in this episode"/research- "research this", "look up", "find information about", "perplexity research"/research-deep- "deep research", "thorough research", "vault-first research", "research gaps"/x-pulse- "x pulse", "what is trending on twitter", "scan x for", "twitter pulse"/x-read- "read this x post", "deep read this tweet", "analyze this tweet", "read this thread"/youtube- "summarize youtube", "youtube transcript", "extract video", "youtube to vault"
Meta - vault setup, health, structure
/create-command- "create command", "new command", "add a command", "scaffold a command"/obsidian-architect- "document this codebase", "architect this project", "map this code into my vault", "generate architecture notes", "refresh architecture docs"/obsidian-export- "export vault", "snapshot vault", "dump vault", "vault export"/obsidian-health- "vault health", "check vault", "audit vault", "vault diagnostics"/obsidian-init- "init vault", "bootstrap vault", "setup vault", "scan vault"/obsidian-visualize- "visualize vault", "vault map", "canvas of vault", "show me the vault shape"
Trigger conventions
Each command file has a leading paragraph that documents its trigger phrases
in plain English. Treat those triggers as primary; the slash-form /name is
also accepted if the user types it.
Scripts
Python helpers live under .codex/scripts/. They run via uv run -m scripts.research.<name> from the vault root (or wherever the skill is
installed). The commands that need them reference the exact invocation
inside the command body.
Generated by adapters/codex-cli/adapter.sh - do not edit manually.