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.
npx agentmods add instructions/magicseek/nblm/agents-mdgit clone --depth 1 https://github.com/magicseek/nblmWhat 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 | $0.01454 | $0.01454 |
| Opus 5 | $0.00727 | $0.00727 |
| Sonnet 5 | $0.00291 | $0.00291 |
| Haiku 4.5 | $0.00145 | $0.00145 |
Grade A, and why
nblm AGENTS.md 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 2d 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.
How it starts
The opening of the file, as written. The whole thing — 162 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
This file provides guidance to AI coding agents when working with code in this repository. It mirrors CLAUDE.md to ensure all AI agents have the same context.
⚠️ IRON RULE - NotebookLM Usage
When working in this repository and needing to reference or query NotebookLM, you MUST use the skill provided by this repo itself. Do not use external NotebookLM tools or services - always use the scripts and tooling defined here.
This is a non-negotiable project law.
Project Overview
nblm - enables AI coding agents to query Google NotebookLM for source-grounded, citation-backed answers. Uses the agent-browser daemon (Node.js) and a Unix socket protocol for automation.
Session Model: Stateless per question; the daemon keeps browser state in memory until it is stopped.
Development Commands
Running Scripts (Always use run.py wrapper)
# CORRECT - Always use run.py:
python scripts/run.py auth_manager.py status
python scripts/run.py notebook_manager.py list
python scripts/run.py ask_question.py --question "..."
# WRONG - Will fail without venv:
python scripts/auth_manager.py status
The run.py wrapper automatically creates .venv, installs Python deps, and installs Node.js deps if needed.
Manual Environment Setup (if automatic fails)
python -m venv .venv
source .venv/bin/activate # Linux/Mac
pip install -r requirements.txt
npm install
npm run install-browsers
Common Script Commands
# Authentication
python scripts/run.py auth_manager.py setup # Default: Google
python scripts/run.py auth_manager.py setup --service zlibrary
python scripts/run.py auth_manager.py status # Show all services
python scripts/run.py auth_manager.py status --service zlibrary
python scripts/run.py auth_manager.py clear --service zlibrary # Clear auth data
# Notebook Library
python scripts/run.py notebook_manager.py list
python scripts/run.py notebook_manager.py add --url URL --name NAME --description DESC --topics TOPICS
python scripts/run.py notebook_manager.py search --query KEYWORD
python scripts/run.py notebook_manager.py activate --id ID
python scripts/run.py notebook_manager.py remove --id ID
# Query
python scripts/run.py ask_question.py --question "..." [--notebook-id ID] [--notebook-url URL] [--show-browser]
# Source Manager
python scripts/run.py source_manager.py add --url "https://zh.zlib.li/book/..."
python scripts/run.py source_manager.py add --file "/path/to/book.pdf"
# Cleanup
python scripts/run.py cleanup_manager.py # Preview
python scripts/run.py cleanup_manager.py --confirm # Execute
python scripts/run.py cleanup_manager.py --preserve-library # Keep notebooks
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.
- 2d ago First seen · 162 lines · 1,454 tokens per session scan A 7ea2b49d8c14
nblm AGENTS.md is an instructions file published in the GitHub repository magicseek/nblm (33 stars, last pushed 6mo ago), licensed MIT. It adds 1,454 tokens to every session, about $0.0073 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 instructions, from other repositories
open-dictionary AGENTS.md
Instructions for ahpxex/open-dictionary, covering open dictionary rewrite charter, product framing, core workflow, technical framework and 1. raw ingestion layer.
designing-real-world-ai-agents-workshop CLAUDE.md
Instructions for iusztinpaul/designing-real-world-ai-agents-workshop, covering project, project structure, tech stack, access documentation and running qa.
ab900 CLAUDE.md
Instructions for timothywarner-org/ab900, covering claude.md, repository purpose, architecture, commands and 2. non-ascii punctuation -- must return zero.
canvas-lms-mcp AGENTS.md
Instructions for bruchris/canvas-lms-mcp, covering agents.md — canvas lms mcp server, quick start, run with npx (no install needed), or install globally and architecture.
obsidian-university-workflow CLAUDE.md
Instructions for ABO896/obsidian-university-workflow, covering obsidian university workflow, project structure, rules for working on this project, templater api — always read the docs first and config alignment.
paper-pilot CLAUDE.md
Instructions for aytzey/paper-pilot, covering claude.md, best fit, setup, first workflow to try and claude prompt patterns.