Skip to content
Back to search

skynet-agent

An Open Source, Claude Code Like Tool, With RAG + Graph RAG + MCP Integration, and Supports Most LLMs (Incomplete But Functional & Usable)Autonomous LLM agent: ChromaDB semantic RAG + Neo4j knowledge-graph sync + MCP tool-call gateway.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today

Check it deeper on your machine

It has changed since its scan on 11 Nov 2026.

As a Claude Code plugin, install it and then ask Claude Code:

/plugin marketplace add keithmackay/mackayi/plugin install scanskillsafety@mackayiDo a deep scan of https://github.com/esinecan/skynet-agent with scanskillsafety

Or from a terminal, clone it and run it:

git clone https://github.com/keithmackay/scanskillsafety-skill ~/.claude/skills/scanskillsafetynode ~/.claude/skills/scanskillsafety/cli.cjs --deep 'https://github.com/esinecan/skynet-agent'

It checks the exact commit: every file, its packages against OSV (api.osv.dev), how it compares with this page's scan (findsafeskills.com receives the repository's address), and, if you agree, Claude reading every file with your own Claude subscription. It never runs the code.

Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/esinecan--skynet-agent/badge.svg)](/listings/esinecan--skynet-agent)

HTML

<a href="/listings/esinecan--skynet-agent"><img src="/api/listings/esinecan--skynet-agent/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

Prerequisites

- Node.js 18+ - Docker & Docker Compose - LLM API key (free Google AI Studio recommended)

Installation

git clone https://github.com/esinecan/skynet-agent.git
cd skynet-agent
npm install

cp .env.example .env.local
# Edit .env.local with your API keys

docker-compose up -d    # ChromaDB (8000) + Neo4j (7474, 7687)
npm run dev             # Or npm run dev:next if Neo4j issues

**Access:** - Application: `http://localhost:3000` - Conscious Memory: `http://localhost:3000/conscious-memory` - Neo4j Browser: `http://localhost:7474` (neo4j/password123)

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

README.md · README · 270 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. # Skynet Agent
  2. > *What if AI could not only access memories, but consciously choose what to remember? With MCP tool access fully supported?*
  3. ![image](https://github.com/user-attachments/assets/0e8d3705-066b-432e-80ae-836e5b75c8ca)
  4. [![TypeScript](https://img.shields.io/badge/TypeScript-007ACC?style=for-the-badge&logo=typescript&logoColor=white)](https://www.typescriptlang.org/)
  5. [![Next.js](https://img.shields.io/badge/Next.js-000000?style=for-the-badge&logo=next.js&logoColor=white)](https://nextjs.org/)
  6. [![ChromaDB](https://img.shields.io/badge/ChromaDB-FF6B6B?style=for-the-badge&logo=database&logoColor=white)](https://www.trychroma.com/)
  7. [![MCP](https://img.shields.io/badge/MCP-4A90E2?style=for-the-badge&logo=protocol&logoColor=white)](https://modelcontextprotocol.io/)
  8. AI conversation platform implementing dual-layer memory architecture inspired by human cognition. Combines automatic background memory with conscious, deliberate memory operations that AI controls. Tool access powers similar to Claude Desktop.
  9. ## Core Features
  10. ### LangGraph-Powered Autopilot
  11. **Purpose-driven autonomous execution** replacing simple query generation with sophisticated multi-step workflows:
  12. - Purpose analysis and strategic planning
  13. - Context gathering from all memory systems
  14. - Smart tool orchestration with error recovery
  15. - Progress monitoring with adaptive replanning
  16. - Reflection engine for continuous learning
  17. - Configurable aggressiveness and safety controls
  18. ### Dual-Layer Memory
  19. **Automatic Memory (RAG)**: Non-volitional background memory using ChromaDB vectors and Google text-embedding-004
  20. **Conscious Memory**: Volitional operations via MCP tools - save, search, update, delete with tags and importance scoring
  21. **Knowledge Graph**: Neo4j-powered relationship mapping with automatic synchronization and retry mechanisms
  22. ### MCP Tool Ecosystem
  23. Exposes memory operations as Model Context Protocol tools for natural conversation flow. Clean separation between UI, memory, and AI operations.
  24. ## Quick Setup
  25. ### Prerequisites
  26. - Node.js 18+
  27. - Docker & Docker Compose
  28. - LLM API key (free Google AI Studio recommended)
  29. ### Installation
  30. ```bash
  31. git clone https://github.com/esinecan/skynet-agent.git
  32. cd skynet-agent
  33. npm install
  34. cp .env.example .env.local
  35. # Edit .env.local with your API keys
  36. docker-compose up -d # ChromaDB (8000) + Neo4j (7474, 7687)
  37. npm run dev # Or npm run dev:next if Neo4j issues
  38. ```
  39. **Access:**
  40. - Application: `http://localhost:3000`
  41. - Conscious Memory: `http://localhost:3000/conscious-memory`
  42. - Neo4j Browser: `http://localhost:7474` (neo4j/password123)
  43. ## Supported LLMs
  44. | Provider | Best For | Model |
  45. |----------|----------|-------|
  46. | Google | Multimodal & speed | `gemini-2.5-flash-preview-05-20` |
  47. | DeepSeek | Cost-effective | `deepseek-chat` |
  48. | OpenAI | Ecosystem | `gpt-4o-mini` |
  49. | Anthropic | Reasoning | `claude-3-5-haiku-20241022` |
  50. | Groq | Ultra-fast | `llama-3.3-70b-versatile` |
  51. | Mistral | Natural language | `mistral-large-latest` |
  52. | Ollama | Privacy | `llama3.2:latest` |
  53. ## Configuration
  54. ### Essential Environment Variables
  55. ```env
  56. # LLM (pick one)
  57. GOOGLE_API_KEY=your_key
  58. DEEPSEEK_API_KEY=your_key
  59. # Main LLM Configuration
  60. LLM_PROVIDER=google
  61. LLM_MODEL=gemini-2.5-flash-preview-05-20
  62. # Motive Force (Autopilot) LLM Configuration (optional - defaults to main LLM)
  63. LLM_PROVIDER_MOTIVE_FORCE=deepseek
  64. LLM_MODEL_MOTIVE_FORCE=deepseek-chat
  65. # Services
  66. CHROMA_URL=http://localhost:8000
  67. NEO4J_URI=bolt://localhost:7687
  68. NEO4J_PASSWORD=password123
  69. # Autopilot
  70. MOTIVE_FORCE_ENABLED=false
  71. MOTIVE_FORCE_MAX_CONSECUTIVE_TURNS=10
  72. MOTIVE_FORCE_TEMPERATURE=0.8
  73. ```
  74. ### Autopilot Usage
  75. Enable via UI toggle. Your next message becomes the objective:
  76. ```
  77. Using timestamps and normal querying, organize today's memories into 5-10 groups.
  78. Delete redundant items, consolidate similar ones, add insights. Check with autopilot
  79. periodically. Daily maintenance cultivates curated memory over time.
  80. ```
  81. Configure via gear icon: turn delays, limits, memory integration, aggressiveness modes.
  82. ## Development
  83. ### Scripts
  84. ```bash
  85. # Development
  86. npm run dev # Full stack + KG sync
  87. npm run dev:debug # With Node debugging
  88. npm run dev:next # Frontend only
  89. npm run dev:kg # KG sync only
  90. # Knowledge Graph
  91. npm run kg:sync # One-time sync
  92. npm run kg:sync:full # Complete resync
  93. npm run kg:sync:queue # Process retry queue
  94. # Testing
  95. npm run test # All tests
  96. npm run test:rag # RAG system
  97. npm run test:neo4j # Neo4j integration
  98. ```
  99. ### Project Structure
  100. ```
  101. skynet-agent/
  102. ├── src/
  103. │ ├── app/ # Next.js routes
  104. │ ├── components/ # React components
  105. │ ├── lib/ # Core libraries
  106. │ │ ├── motive-force-graph.ts # LangGraph workflow
  107. │ │ ├── conscious-memory.ts # Volitional memory
  108. │ │ ├── rag.ts # Automatic memory
  109. │ │ └── knowledge-graph-*.ts # Neo4j integration
  110. │ └── types/ # TypeScript definitions
  111. ├── docker-compose.yml # Services setup
  112. └── motive-force-prompt.md # Autopilot personality
  113. ```
  114. ## Memory Architecture
  115. ### Automatic Memory (RAG)
  116. ```typescript
  117. interface Memory {
  118. id: string;
  119. text: string;
  120. embedding: number[]; // Google text-embedding-004
  121. metadata: {
  122. sender: 'user' | 'assistant';
  123. timestamp: string;
  124. summary?: string; // Auto-summarized if over limit
  125. };
  126. }
  127. ```
  128. ### Conscious Memory
  129. ```typescript
  130. interface ConsciousMemory {
  131. id: string;
  132. content: string;
  133. tags: string[];
  134. importance: number; // 1-10
  135. source: 'explicit' | 'suggested' | 'derived';
  136. metadata: {
  137. accessCount: number;
  138. lastAccessed: string;
  139. };
  140. }
  141. ```
  142. ### LangGraph State
  143. ```typescript
  144. interface MotiveForceGraphState {
  145. messages: BaseMessage[];
  146. currentPurpose: string;
  147. subgoals: SubGoal[];
  148. executionPlan: ExecutionStep[];
  149. toolResults: ToolResult[];
  150. reflections: Reflection[];
  151. overallProgress: number;
  152. blockers: string[];
  153. needsUserInput: boolean;
  154. }
  155. ```
  156. ## API Reference
  157. ### Conscious Memory
  158. ```http
  159. POST /api/conscious-memory
  160. {
  161. "action": "save|search|update|delete|stats|tags",
  162. "content": "string",
  163. "tags": ["array"],
  164. "importance": 7
  165. }
  166. ```
  167. ### Autopilot
  168. ```http
  169. POST /api/motive-force
  170. {
  171. "action": "generate|generateStreaming|saveConfig|getState",
  172. "sessionId": "string",
  173. "data": {}
  174. }
  175. ```
  176. ## Advanced Features
  177. ### Hybrid Search
  178. 1. **Semantic**: Vector similarity via embeddings
  179. 2. **Keyword**: Exact match fallback
  180. 3. **Smart Merge**: Intelligent ranking with deduplication
  181. ### Knowledge Graph Sync
  182. - Automatic extraction from chat history
  183. - Background service with retry queue
  184. - Metrics collection and error handling
  185. - Eventually consistent with ChromaDB
  186. ### Safety Mechanisms
  187. - Turn limits and error counting
  188. - Manual override capabilities
  189. - Resource usage monitoring
  190. - Emergency stop functionality
  191. ## Troubleshooting
  192. **"Embeddings service unavailable"**: Falls back to hash-based embeddings. Check Google API key.
  193. **"ChromaDB connection failed"**: Ensure `docker-compose up -d` and port 8000 available.
  194. **"Neo4j sync errors"**: Check credentials, run `npm run kg:sync:queue` for retries.
  195. **"Actually Looks Very Ugly"**: I suck at UI design.
  196. ## Development Philosophy
  197. Inspired by cognitive science:
  198. - **Dual-Process Theory**: Automatic vs controlled processes
  199. - **Memory Consolidation**: Active organization
  200. - **Working Memory**: Conscious manipulation
  201. Technical innovations:
  202. - **Hybrid Search**: Solves subset query limitations
  203. - **MCP Architecture**: Natural language memory control
  204. - **Importance Weighting**: Smart prioritization
  205. - **LangGraph Integration**: Complex autonomous workflows
  206. ## Contributing
  207. Fork, improve, PR. Areas: memory algorithms, UI/UX, MCP tools, autopilot intelligence, testing, performance.
  208. ## License
  209. MIT - Lok Tar Ogar!
  210. ## Acknowledgments
  211. ChromaDB, Google AI, Anthropic MCP, Next.js, Neo4j teams. Open source MCP servers and Ollama Vercel AI SDK library.

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.