$ show Fosowl/agenticSeek
AgenticSeek is a private, local autonomous AI assistant built as an alternative to Manus AI. The system features a multi-agent architecture with six specialized agents (Casual, Coder, File, Browser, Planner, MCP) selected by a hybrid router using BART zero-shot classification and an adaptive few-shot classifier. It provides a React frontend, FastAPI backend with Celery, CLI with voice support, and Docker Compose orchestration for self-hosted services including SearxNG search and Redis caching. The codebase demonstrates solid engineering with clean abstractions for LLM providers (14+ options), tool execution, browser automation, and workspace isolation. It is under active development but shows signs of technical drift in deprecated components (llm_server, SerpApi-based web search) and fragile browser automation. The architecture is modular and contains several reusable patterns, though the codebase is tightly coupled to its specific agent system and has substantial external dependencies.
The LLM identified 4 files worth highlighting for reuse:
Abstract base agent class that implements a robust execution loop: LLM request parsing, tool call extraction via block detection, tool execution, and memory management. Demonstrates a reusable pattern for building autonomous agents with configurable tools and system prompts.
Hybrid router that combines BART zero-shot classification with an AdaptiveClassifier few-shot learner to select the optimal agent for a given query. Includes voting logic and confidence scoring, offering a reusable pattern for dynamic task routing in multi-agent systems.
Base tool class with a clean interface for execution, error checking, and feedback. Includes a reusable `load_exec_block()` method to parse code blocks from LLM output, supporting extensibility for adding new tools with consistent error handling.
Workspace manager that resolves file paths securely, detecting and preventing path traversal attacks. Provides separate runtime and work directories, demonstrating a reusable pattern for sandboxed file system access in agent systems.