</>MCP Agents Market
Agent

GenericAgent

by lsdefine14kPythonUpdated 2026-08-24

Self-evolving agent: grows skill tree from 3.3K-line seed, achieving full system control with 6x less token consumption

GenericAgent is a minimal, self-evolving autonomous AI agent framework built on approximately 3,300 lines of core code. It grants LLMs system-level control over local computers through 9 atomic tools covering browser automation, terminal access, file operations, keyboard/mouse input, screen vision, and mobile device control via ADB. The agent automatically crystallizes each task solution into reusable Skills that accumulate over time, forming a personal skill tree grown from the seed codebase. It achieves 6× lower token consumption than competing agents while maintaining high task completion rates across web automation, system control, and development tasks.

Key Features

Self-evolution mechanism that automatically converts task executions into reusable Skills stored in hierarchical memory (L0-L4 layers)
Minimal architecture with ~3K lines core code, ~100-line agent loop, and only 9 atomic tools (code_run, file operations, web control, vision, user interaction)
Real browser control via TMWebdriver that injects into live Chrome sessions, preserving login state and passing bot detection (56/56 SannySoft tests)
Token-efficient design using <30K context window versus 200K-1M for other agents, reducing costs and hallucinations
Cross-platform system control including keyboard/mouse automation, screen vision, mobile ADB control, and terminal access
Multi-model compatibility supporting Claude, Gemini, Kimi, MiniMax and other major LLMs
Multiple frontend options including terminal UI, Streamlit web UI, desktop GUI, and IM integrations (Telegram, Discord, WeChat, Lark)
Self-bootstrap proof: entire repository was created autonomously by GenericAgent from git init to commit messages

Use Cases

  • 01Web automation tasks that require persistent login sessions and bot-detection evasion (e.g., passing hCaptcha challenges during Discord bot setup)
  • 02Food delivery ordering and e-commerce automation by navigating apps, selecting items, and completing checkout flows
  • 03Financial workflows including quantitative stock screening with technical indicators and expense tracking via mobile app control
  • 04Batch messaging and communication automation across WeChat, Telegram, and other IM platforms
  • 05Development environment setup and code repository management with autonomous Git operations and dependency installation
  • 06Long-running autonomous tasks like periodic web content summarization and monitoring with scheduled execution

Related Agents

View more

GenericAgent — FAQ

What is GenericAgent?+

GenericAgent is a self-evolving autonomous AI agent framework that uses 9 atomic tools and a minimal codebase to give LLMs full system control over computers, browsers, and mobile devices. It automatically learns from each task by crystallizing execution paths into reusable Skills that grow over time.

How do I install GenericAgent?+

Clone the repository, create a virtual environment with Python 3.11 or 3.12 (not 3.14), install with 'uv pip install -e ".[ui]"', copy mykey_template.py to mykey.py, and add your LLM API keys. Alternatively, use the one-line installer script for Windows (PowerShell) or Linux/macOS that sets up an isolated environment.

Which LLM providers does GenericAgent work with?+

GenericAgent supports Claude (Sonnet/Opus), Gemini, Kimi, MiniMax, and other major LLM providers. You configure API keys in the mykey.py file and can switch between models based on task requirements.

What are the prerequisites and do I need API keys?+

You need Python 3.11 or 3.12, and API keys for at least one supported LLM provider (Claude, Gemini, etc.). For web automation, you manually drag the included Chrome extension into chrome://extensions; other capabilities like OCR and vision are set up by instructing the agent itself.

Is GenericAgent free to use?+

GenericAgent is open-source under the MIT License and free to use. You only pay for the LLM API costs from your chosen provider, though token consumption is significantly lower (6×) than competing agent frameworks.

How does the self-evolution mechanism work?+

Each time GenericAgent completes a task, it automatically extracts the execution path and stores it as a Skill in its layered memory system (L0-L4). On subsequent similar tasks, it recalls and reuses these Skills directly rather than re-exploring, creating a growing personal skill tree unique to your usage patterns.

How do I install GenericAgent?+

Open the source repository on GitHub and follow its README. GenericAgent is a agent — MCP Agents Market links you directly to the official repo.

Is GenericAgent free?+

GenericAgent is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.

Related searches