Mirofish Offline Simulation
Run fully local Neo4j plus Ollama multi-agent simulations to stress-test how documents might shift public opinion or market sentiment.
Overview
Mirofish-offline-simulation is an agent skill most often used in Build (also Validate) that installs and runs a local Neo4j and Ollama multi-agent swarm to simulate reactions to documents.
Install
npx skills add https://github.com/aradotso/trending-skills --skill mirofish-offline-simulationWhat is this skill?
- End-to-end offline pipeline: document in → NER/graph build → persona swarm → hourly social dynamics → report output
- Neo4j CE 5.15 knowledge graph with nomic-embed-text embeddings—no cloud LLM APIs required
- Ollama-served LLMs for entity extraction, agent posts, replies, and opinion shifts
- Triggers cover install, configure, and run public-reaction or market-sentiment simulations
- ReportAgent-style interview summaries for focus narratives after swarm runs
- Neo4j Community Edition 5.15 specified in skill
- Hundreds of generated agent personas in simulation env setup
Adoption & trust: 1.2k installs on skills.sh; 31 GitHub stars; 1/3 security scanners passed (skills.sh audits).
What problem does it solve?
You need to preview public or market sentiment shifts from a document without sending sensitive text to cloud LLM APIs.
Who is it for?
Indie builders prototyping sentiment labs, policy sandboxes, or launch comms drills where local LLMs and graph memory are acceptable tradeoffs for latency and ops.
Skip if: Teams that need production-grade forecast accuracy, minimal DevOps, or instant results without Neo4j and Ollama maintenance.
When should I use this skill?
Triggers include set up mirofish offline simulation, run multi-agent social simulation locally, simulate public reaction to a document, configure neo4j ollama agent simulation, or simulate market sentiment with local llm
What do I get? / Deliverables
You get a configured offline simulation path that produces agent-generated posts, arguments, and summary reports from your source document.
- Configured MiroFish-Offline environment
- Simulation run with agent posts and opinion dynamics
- Summary or report output from simulation agents
Recommended Skills
Journey fit
Spans multiple journey phases - primary shelf plus alternate fits below.
Canonical shelf is Build because installation wires Neo4j, Ollama, and agent personas into your product stack. Integrations subphase matches connecting graph DB, local LLMs, and simulation runtime—not a one-off script.
Where it fits
Simulate social pushback to a landing-page promise before you commit to full build.
Wire Neo4j and Ollama services into your repo following the skill’s install and configure steps.
Rehearse announcement copy against synthetic audience threads to refine launch messaging.
How it compares
Skill-orchestrated local simulation stack, not a hosted sentiment API or single-shot chat prompt.
Common Questions / FAQ
Who is mirofish-offline-simulation for?
Solo developers and small teams adding offline multi-agent social or market-sentiment simulations to apps using Neo4j and Ollama.
When should I use mirofish-offline-simulation?
During Validate prototype runs to rehearse messaging reactions, during Build when integrating graph plus local LLM agents, and before Launch distribution when testing narrative risk in private environments.
Is mirofish-offline-simulation safe to install?
Check this page’s Security Audits panel and audit shell/network steps before running on machines with sensitive documents or open ports.
SKILL.md
READMESKILL.md - Mirofish Offline Simulation
# MiroFish-Offline Skill > Skill by [ara.so](https://ara.so) — Daily 2026 Skills collection. MiroFish-Offline is a fully local multi-agent swarm intelligence engine. Feed it any document (press release, policy draft, financial report) and it generates hundreds of AI agents with unique personalities that simulate public reaction on social media — posts, arguments, opinion shifts — hour by hour. No cloud APIs required: Neo4j CE 5.15 handles graph memory, Ollama serves the LLMs. --- ## Architecture Overview ``` Document Input │ ▼ Graph Build (NER + relationship extraction via Ollama LLM) │ ▼ Neo4j Knowledge Graph (entities, relations, embeddings via nomic-embed-text) │ ▼ Env Setup (generate hundreds of agent personas with personalities + memory) │ ▼ Simulation (agents post, reply, argue, shift opinions on simulated platforms) │ ▼ Report (ReportAgent interviews focus group, queries graph, generates analysis) │ ▼ Interaction (chat with any individual agent, full memory persists) ``` **Backend**: Flask + Python 3.11 **Frontend**: Vue 3 + Node 18 **Graph DB**: Neo4j CE 5.15 (bolt protocol) **LLM**: Ollama (OpenAI-compatible `/v1` endpoint) **Embeddings**: `nomic-embed-text` (768-dimensional, via Ollama) **Search**: Hybrid — 0.7 × vector similarity + 0.3 × BM25 --- ## Installation ### Option A: Docker (Recommended) ```bash git clone https://github.com/nikmcfly/MiroFish-Offline.git cd MiroFish-Offline cp .env.example .env # Start Neo4j + Ollama + MiroFish backend + frontend docker compose up -d # Pull required models into the Ollama container docker exec mirofish-ollama ollama pull qwen2.5:32b docker exec mirofish-ollama ollama pull nomic-embed-text # Check all services are healthy docker compose ps ``` Open `http://localhost:3000`. ### Option B: Manual Setup **1. Neo4j** ```bash docker run -d --name neo4j \ -p 7474:7474 -p 7687:7687 \ -e NEO4J_AUTH=neo4j/mirofish \ neo4j:5.15-community ``` **2. Ollama** ```bash ollama serve & ollama pull qwen2.5:32b # Main LLM (~20GB, requires 24GB VRAM) ollama pull qwen2.5:14b # Lighter option (~10GB VRAM) ollama pull nomic-embed-text # Embeddings (small, fast) ``` **3. Backend** ```bash cp .env.example .env # Edit .env (see Configuration section) cd backend pip install -r requirements.txt python run.py # Backend starts on http://localhost:5000 ``` **4. Frontend** ```bash cd frontend npm install npm run dev # Frontend starts on http://localhost:3000 ``` --- ## Configuration (`.env`) ```bash # ── LLM (Ollama OpenAI-compatible endpoint) ────────────────────────── LLM_API_KEY=ollama LLM_BASE_URL=http://localhost:11434/v1 LLM_MODEL_NAME=qwen2.5:32b # ── Neo4j ───────────────────────────────────────────────────────────── NEO4J_URI=bolt://localhost:7687 NEO4J_USER=neo4j NEO4J_PASSWORD=mirofish # ── Embeddings (Ollama) ─────────────────────────────────────────────── EMBEDDING_MODEL=nomic-embed-text EMBEDDING_BASE_URL=http://localhost:11434 # ── Optional: swap Ollama for any OpenAI-compatible provider ───────── # LLM_API_KEY=$OPENAI_API_KEY # LLM_BASE_URL=https://api.openai.com/v1 # LLM_MODEL_NAME=gpt-4o ``` --- ## Core Python API ### GraphStorage Interface The abstraction layer between MiroFish and the graph database: ```python from backend.storage.base import GraphStorage from backend.storage.neo4j_storage import Neo4jStorage # I