obsidian-second-brain turns an Obsidian vault into persistent, searchable memory for Claude Code and other command-line coding agents, storing knowledge as linked Markdown notes. It is for developers, founders, writers, and researchers who want agents to retain project context across sessions. Its catalogue entries provide commands, hooks, a plugin, a skill, and instructions for capturing, finding, and maintaining that memory.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/eugeniughelbur/obsidian-second-brainWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/eugeniughelbur/obsidian-second-brain/notebooklm)<a href="https://agentmods.dev/commands/eugeniughelbur/obsidian-second-brain/notebooklm"><img src="https://agentmods.dev/badge/commands/eugeniughelbur/obsidian-second-brain/notebooklm.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00035 | $0.01120 |
| Opus 5 | $0.00017 | $0.00560 |
| Sonnet 5 | $0.00007 | $0.00224 |
| Haiku 4.5 | $0.00003 | $0.00112 |
Grade A, and why
notebooklm scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
Use the obsidian-second-brain skill. Execute /notebooklm [topic]:
-
Resolve the topic from the user's argument. If no topic, ask: "What topic for source-grounded research?"
-
Run the script from the skill root (its absolute path was given at session start as Skill root; substitute it for
SKILL_ROOT):uv run --directory "SKILL_ROOT" -m scripts.research.notebooklm --topic "<topic>" -
The script does the whole flow end-to-end:
- Scans the vault for the top 12 relevant notes (same shape as
/research-deepPhase 1). - Uploads them to a fresh Gemini File Search store.
- Asks Gemini (default
gemini-2.5-flash, override viaNOTEBOOKLM_MODELenv) for a synthesis grounded against those sources. - Writes the AI-first synthesis to
Research/NotebookLM/YYYY-MM-DD - <slug>.md. - Deletes the File Search store so nothing is left behind.
- Emits a
<<<NOTEBOOKLM_PROPAGATION_PAYLOAD>>>JSON block.
- Scans the vault for the top 12 relevant notes (same shape as
-
After save, do the propagation step. Same flow as
/research-deep:- Parse the propagation payload.
- Read the saved synthesis at
saved_note. - Treat the synthesis as the "conversation context" input to
/obsidian-save. - Run the standard
/obsidian-saveflow: spawn parallel subagents (People, Projects, Tasks, Decisions, Ideas) and update vault notes per any "Recommended next reads or angles" bullets if they map to entities or projects. - Link the new synthesis note from today's daily note.
-
Report back to the user: "Saved [[YYYY-MM-DD - ]] to Research/NotebookLM/. Linked from today's daily note. Updated [[X]], created [[Y]]."
-
Plain English triggers: "notebooklm this", "ground research on X using my vault", "source-grounded research on X", "ask my own notes about X".
-
When to choose
/notebooklmover/research-deep:/research-deep(Perplexity + Grok): when you want OPEN-WEB + X-discourse coverage. Cost: $0.20-0.80./notebooklm(Gemini File Search): when you want answers GROUNDED IN your own vault. Cost: ~$0.01-0.05.- Run both for high-value topics. The web view and the grounded view rarely contradict, and the contradictions are where the insight is.
-
Configuration: requires
GEMINI_API_KEYin~/.config/obsidian-second-brain/.env. Get one free at https://aistudio.google.com/apikey. OptionalNOTEBOOKLM_MODELoverride (defaultgemini-2.5-flash).
AI-first rule: Every note created or updated by this command MUST follow references/ai-first-rules.md. If that path does not resolve from your working directory, search upward for it; if you still cannot read it, say so before writing rather than producing a note that silently skips the rule. The saved synthesis at Research/NotebookLM/YYYY-MM-DD - <slug>.md follows the template baked into the script (preamble, frontmatter, vault-baseline links, response verbatim). Do not strip those.
Anti-fabrication: Search exhaustively before claiming any note, person, or file is absent - false absence is the most common failure mode - and never invent facts, entities, or dates (mark unknowns as TBD). See the anti-fabrication and search-completeness hard rules in references/ai-first-rules.md.
Why Gemini File Search and not the browser: NotebookLM has no public API for personal Google accounts. Gemini File Search (generally available, plain API key, same Gemini model family) gives the same architectural shape: source-grounded retrieval, multi-document context, citation-style synthesis. One HTTP call, no manual paste step.
Cost: $0.15 per million tokens indexed, storage free, generation at standard Gemini token rates. For a 12-note vault bundle (~30K tokens), expect $0.01-0.05 per run.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 9d ago First seen · 54 lines · 35 tokens per session scan A 2e21c85a80ad
notebooklm is a command published in the GitHub repository eugeniughelbur/obsidian-second-brain (4,364 stars, last pushed yesterday), licensed MIT. It adds 35 tokens to every session and 1,120 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other commands, from other repositories
nano-banana
Generate or edit images using Google's Gemini Image model.
cierre
A sales call just ended: turn its transcript into the full follow-up (CRM, tasks, email draft, reminder, coaching).
daily-journal
Daily journal interview and entry creator with emotional floor tagging.
deconstruct
First-principles analyst: surface hidden assumptions, find foundational truths, rebuild from scratch.
diagnose
Run a self-check on your AI Brain Starter install (CLAUDE.md, Meta folder, skills, hooks, MCPs).
evolve
Cluster hardened instincts into a proposed Command / Skill / Agent (Instinct Engine).