agents

# Tool List AI Agent monorepo — Autonomous Intelligent System Collection, covering macOS engineering execution, browser automation, and knowledge graph context engine. ## Projects --- ### [agent-macos](agent-macos/) — macOS Engineering Execution System MacEngineer AgentOS, a macOS-prioritized engineering execution intelligent system. Based on Claude Code, with four layers of dynamic context injection, risk-grading command strategy engine, JSONL audit log, and Ralph continuous autonomy loop engine. **Core Capabilities**: Hierarchical Context Memory | Strategic Routing Execution | Full-Chain Audit | Cyclic Autonomy Engine ``` Runtime: Bun + Node.js (TypeScript) ``` > Click [agent-macos/](agent-macos/) to view the full documentation --- ### [agent-browser](agent-browser/) — Local Browser Intelligent Agent Framework LZ Browser Agent, a browser automation intelligent agent framework for local and privatized scenarios. Based on browser-use, supports natural language-driven web interaction (click, navigation, data extraction), and can access Ollama local large model or OpenAI Compatible API. **Core Capabilities**: Natural Language-Driven Browser | Multi-LLM Adaptation | MCP Service | Local Priority Deployment ``` Runtime: Python 3.11+ / uv ``` > Click [agent-browser/](agent-browser/) to view the full documentation --- ### [agent-graphiti](agent-graphiti/) — Temporal Knowledge Graph Engine Graphiti Custom, a temporal context graph engine for AI Agents. Continuously writes conversations, documents, and business events into graph structures, supporting joint retrieval by time, entity relationships, and semantic similarity. Compatible with Neo4j, FalkorDB, Kuzu, and Neptune graph database backends. **Core Capabilities**: Temporal Fact Storage | Episode Tracing | Hybrid Retrieval | MCP Server ``` Runtime: Python 3.10+ / uv Graph DB: Neo4j | FalkorDB | Kuzu | Neptune ``` > Click [agent-graphiti/](agent-graphiti/) to view the full documentation --- ## Quick Start Each subproject is self-contained with its own dependencies and build system. ```bash # agent-macos cd agent-macos && pnpm install && pnpm run dev # agent-browser cd agent-browser && uv sync && uv run python -m browser_use # agent-graphiti cd agent-graphiti && uv sync && uv run python -m graphiti_core ``` ## Repository Structure ``` agents/ ├── agent-macos/ # TypeScript/Bun — macOS Engineering Execution Intelligent Agent ├── agent-browser/ # Python — Browser Automation Intelligent Agent ├── agent-graphiti/ # Python — Temporal Knowledge Graph Engine ├── .github/workflows/ # Unified CI (triggers by subproject path) └── README.md ``` ## CI/CD Unified CI with path-based triggers — only runs checks for changed subprojects. ## License Each sub-project maintains its own license.