neurostack CLAUDE.md

Project instructions for NeuroStack, a local knowledge-management system that stores Markdown notes, indexes them for search, and offers a command-line tool, an MCP server, and an OpenAI-compatible API. MCP is a standard way for AI assistants to discover and use external tools.

In plain words
What is it for?
Use them to install and initialize NeuroStack, connect it to Claude Code through its MCP server, run its local API, and search or retrieve information from a Markdown vault.
Why use it?
They give an AI assistant the setup commands, available interfaces, and search options needed to work with the project. They also explain how to run the system locally and configure authentication for its API.

Instructions file

Install

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.

agentmods
npx agentmods add instructions/raphasouthall/neurostack/claude-md
Clone the repo
git clone --depth 1 https://github.com/raphasouthall/neurostack
Per session 3,464 This file is loaded in full into every session.
When invoked 3,464 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.03464 $0.03464
Opus 5 $0.01732 $0.01732
Sonnet 5 $0.00693 $0.00693
Haiku 4.5 $0.00346 $0.00346

Measured 2d ago against content hash c8e6baaf4945, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

neurostack CLAUDE.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.

CLAUDE.md · 238 lines

How it starts

The opening of the file, as written. The whole thing — 238 lines — stays where its author put it; the contents beside it link to each section on GitHub.

NeuroStack - Claude Code Guide

NeuroStack is a neuroscience-grounded knowledge management system. CLI + MCP server + OpenAI-compatible API. Everything runs locally against a Markdown vault indexed in SQLite + FTS5.

Quick Reference

Installation

npm install -g neurostack    # bootstraps CLI, Python, uv, deps
neurostack init              # lite/full → vault setup → index

MCP Server (recommended for Claude Code)

Add to ~/.claude/.mcp.json:

{
  "mcpServers": {
    "neurostack": {
      "command": "neurostack",
      "args": ["serve"],
      "env": {}
    }
  }
}

OpenAI-compatible API

pip install neurostack[api]
neurostack api                          # localhost:8000
neurostack api --host 0.0.0.0 --port 9000
NEUROSTACK_API_KEY=secret neurostack api  # with auth

Endpoints: POST /v1/chat/completions, POST /v1/embeddings, GET /v1/models, GET /health Models: neurostack-ask (RAG), neurostack-search (hybrid), neurostack-tiered (auto-depth), neurostack-triples (facts)

CLI Commands

Search & Retrieval

Command Description
neurostack search "query" Hybrid FTS5 + semantic search. Flags: --mode hybrid|semantic|keyword, --top-k N, --context "domain", --workspace "path/"
neurostack ask "question" RAG Q&A with inline [[citations]]. Uses Ollama LLM. Flags: --top-k N, --workspace
neurostack tiered "query" Token-efficient tiered retrieval. --depth triples|summaries|full|auto
neurostack triples "query" Search knowledge graph triples (subject-predicate-object facts)
neurostack summary "note.md" Get pre-computed AI summary of a note
neurostack graph "note.md" Wiki-link neighborhood with PageRank scores. --depth N
neurostack graph-analysis Structural gaps (related but unlinked) + bridge notes. --top-k N, --min-shared N
neurostack diff Change feed since a baseline or date. --since DATE, --baseline NAME, --checkpoint
neurostack related "note.md" Find semantically similar notes by embedding distance
neurostack communities query "question" GraphRAG global queries across topic clusters
neurostack context "task description" Assemble task-scoped context for session recovery
neurostack brief Compact session briefing (recent activity, hot notes, alerts)

Read the full file on GitHub · 238 lines

Changes

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.

  1. 2d ago First seen · 238 lines · 3,464 tokens per session scan A c8e6baaf4945

Subscribe to this mod's changes

neurostack CLAUDE.md is an instructions file published in the GitHub repository raphasouthall/neurostack (46 stars, last pushed 3d ago), licensed Apache-2.0. It adds 3,464 tokens to every session, about $0.0173 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.

Related

Other instructions, from other repositories