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/jupyter-ai-contrib/nb-cli/agents-mdgit clone --depth 1 https://github.com/jupyter-ai-contrib/nb-cliWrote 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/instructions/jupyter-ai-contrib/nb-cli/agents-md)<a href="https://agentmods.dev/instructions/jupyter-ai-contrib/nb-cli/agents-md"><img src="https://agentmods.dev/badge/instructions/jupyter-ai-contrib/nb-cli/agents-md.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 | $0.01034 | $0.01034 |
| Opus 5 | $0.00517 | $0.00517 |
| Sonnet 5 | $0.00207 | $0.00207 |
| Haiku 4.5 | $0.00103 | $0.00103 |
Grade A, and why
nb-cli 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 6d 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 — 70 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Agent Guidelines
Working with Notebooks (.ipynb files)
When the user asks to read, edit, execute, or work with .ipynb files, use the notebook-cli skill, which provides the nb command-line tool. Do not use the built-in Read/Write tools for .ipynb files.
Connect-mode integration tests: backend selection
tests/integration_connect_mode.rs exercises connect-mode against whatever
collaboration backend is installed in the active test venv. jupyter-collaboration
and jupyter-server-documents (JSD) are competing collaborative-editing server
extensions and must never be installed into the same venv — each has its own:
tests/.test-venv— JSD + local-mode tests (default). Pinned:jupyter_server==2.20.0,jupyter-server-documents==0.2.5.tests/.test-venv-collab— jupyter-collaboration. Pinned:jupyter_server==2.20.0,jupyter-collaboration==4.4.1.
Set up a venv with ./tests/setup_test_env.sh [jsd|jupyter-collaboration]
(defaults to jsd). Select which backend a test run targets with
NB_TEST_BACKEND=<jsd|jupyter-collaboration> (read by test_helpers::test_backend());
this also picks the matching venv directory automatically. Run with:
NB_TEST_BACKEND=jupyter-collaboration cargo test --test integration_connect_mode -- --test-threads=1
The shared Jupyter server is spawned once per test process (OnceLock) with its
current_dir set to a tempdir root, so backend-specific artifacts like
jupyter-collaboration's .jupyter_ystore.db land there instead of the crate
root. On teardown, an atexit hook calls jupyter server stop <port> -y to
cleanly shut down the server. Each notebook-executing test also explicitly
deletes its Jupyter session/kernel via DELETE /api/sessions/{id} when its
NotebookSession guard drops (production code intentionally never deletes
sessions, so tests must do this themselves).
Known state (2026-07-05): against jupyter-collaboration, the 4
execute/restart tests (what PR #99 / issue #92 fixed — FileID fallback, sessionId
on the Y.js room WS handshake, v1 kernel-WS subprotocol, client-side output
writing) pass 10/10 runs with zero flakiness, and gate the test-connect-collab
CI job. test_clear_outputs_in_connect_mode and
test_clear_outputs_specific_cell_in_connect_mode are marked #[ignore]
against jupyter-collaboration (issue #100): nb output clear correctly edits
the Y.js room, but jupyter_server_ydoc only flushes the room to disk on a
debounced ~1s timer (document_save_delay), so nb read immediately
afterward races that debounce and can observe stale content — confirmed by
direct measurement (still stale at +0.7s, cleared by +1.7s). This is a
different root cause from #90 (JSD's clear never persists, permanently,
because externalized output files get unconditionally re-materialized into
the notebook on every save) — don't conflate the two if either gets fixed.
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.
- 6d ago First seen · 70 lines · 1,034 tokens per session scan A d72754ebd733
nb-cli AGENTS.md is an instructions file published in the GitHub repository jupyter-ai-contrib/nb-cli (90 stars, last pushed 1mo ago), licensed BSD-3-Clause. It adds 1,034 tokens to every session, about $0.0052 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
ai-engineering-labs CLAUDE.md
Instructions for maxencebernardhub/ai-engineering-labs, a project described as: Hands-on AI engineering labs: OpenAI, Anthropic, Google, local models. From a first API call to an agent deployed live on AWS: RAG, LangChain, LangGraph & Deep Agents, multi-provider routing, offline inference, FastAPI, Docker, Lambda.
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.
ab900 CLAUDE.md
Instructions for timothywarner-org/ab900, covering claude.md, repository purpose, architecture, commands and 2. non-ascii punctuation -- must return zero.
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.
agentic-ai-engineering-course AGENTS.md
AGENTS.md instructions for towardsai/agentic-ai-engineering-course, covering the what and course map (lessons ↔ code projects).
anki-mcp-server CLAUDE.md
Claude Code instructions for nailuoGG/anki-mcp-server, covering claude.md, repository overview, development commands, build & development and install dependencies.