agent-code 2.0.0__tar.gz → 2.1.0__tar.gz

agent_code-2.1.0/PKG-INFO

@@ -0,0 +1,1048 @@
+Metadata-Version: 2.4
+Name: agent-code
+Version: 2.1.0
+Summary: Local AI agent with optional browser automation, sandboxed shell, and plug-and-play skills
+Author: minecraftbefile-maker
+License-Expression: MIT
+Project-URL: Homepage, https://agent-code-dev.github.io
+Project-URL: Repository, https://github.com/Agent-code-dev/Agent-code
+Project-URL: Issues, https://github.com/Agent-code-dev/Agent-code/issues
+Project-URL: Changelog, https://github.com/Agent-code-dev/Agent-code/releases
+Keywords: ai,agent,llm,browser,automation,skills,openai
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Environment :: Console
+Classifier: Intended Audience :: Developers
+Classifier: Operating System :: OS Independent
+Classifier: Operating System :: Microsoft :: Windows
+Classifier: Operating System :: POSIX :: Linux
+Classifier: Operating System :: MacOS
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: 3.14
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: openai>=1.0.0
+Requires-Dist: customtkinter>=5.2.2
+Requires-Dist: markdown-it-py>=3.0.0
+Requires-Dist: mcp<2,>=1.9
+Requires-Dist: keyring>=24.2
+Requires-Dist: paramiko>=3.4.0
+Requires-Dist: requests>=2.31.0
+Requires-Dist: DrissionPage>=4.1.1.4
+Requires-Dist: selenium>=4.20.0
+Dynamic: license-file
+
+<p align="center">
+  <img src="src/agent_code/data/icons/agent-code.png" width="145" alt="Agent Code Logo">
+</p><h1 align="center">Agent Code 2.1</h1><h3 align="center">⚡ Your AI coding agent. Your tools. Your environment.</h3><p align="center">
+  An open-source, extensible AI coding agent built with Python and CustomTkinter.
+  <br>
+  Build, modify, debug, and manage projects using AI — with customizable Skills, application Add-ons, and MCP integrations.
+</p><p align="center">
+  <a href="https://pypi.org/project/agent-code/">
+    <img src="https://img.shields.io/pypi/v/agent-code?style=for-the-badge&logo=pypi&logoColor=white" alt="PyPI Version">
+  </a>
+  <a href="https://www.python.org/">
+    <img src="https://img.shields.io/badge/Python-3.10%2B-3776AB?style=for-the-badge&logo=python&logoColor=white" alt="Python 3.10+">
+  </a>
+  <a href="LICENSE">
+    <img src="https://img.shields.io/badge/Application-MIT-2ea44f?style=for-the-badge" alt="MIT License">
+  </a>
+  <a href="server/LICENSE">
+    <img src="https://img.shields.io/badge/Server-AGPL--3.0-blue?style=for-the-badge" alt="Server AGPL License">
+  </a>
+</p><p align="center">
+  <a href="https://github.com/agent-code-dev/agent-code">
+    <img src="https://img.shields.io/badge/GitHub-Repository-181717?style=flat-square&logo=github" alt="GitHub">
+  </a>
+  <a href="https://pypi.org/project/agent-code/">
+    <img src="https://img.shields.io/badge/PyPI-Install-3775A9?style=flat-square&logo=pypi" alt="PyPI">
+  </a>
+  <a href="https://agent-code-dev.github.io">
+    <img src="https://img.shields.io/badge/Website-Visit-4285F4?style=flat-square&logo=googlechrome&logoColor=white" alt="Website">
+  </a>
+</p><p align="center">
+  <b>Windows · Linux · macOS</b>
+  <br>
+  <sub>Python 3.10+ · OpenAI-compatible AI providers · Extensible architecture</sub>
+</p>---
+
+🌟 What is Agent Code?
+
+Agent Code is an open-source desktop AI coding agent designed to make AI-assisted software development more extensible and customizable.
+
+It combines an AI-powered development workspace with a flexible tool architecture, allowing developers to expand both the capabilities of the AI agent and the application itself.
+
+Instead of locking users into one provider or one fixed collection of tools, Agent Code gives developers control over their AI models, integrations, execution environments, and extensions.
+
+🚀 The idea behind Agent Code
+
+- 🤖 Use different AI providers with your own API keys.
+- 🧩 Give the AI additional capabilities through native Python Skills.
+- 🔌 Extend the desktop application through Add-ons.
+- 🌐 Connect external tools through MCP.
+- 💻 Work with local projects and development environments.
+- 📦 Use an optional Alpine Linux sandbox.
+- 🖥️ Connect to remote development machines through SSH.
+- ⚙️ Customize agent workflows and execution permissions.
+
+One application. Multiple extension systems. Your choice of AI provider.
+
+---
+
+✨ Features
+
+<table>
+  <tr>
+    <td width="50%" valign="top">🤖 AI Coding Agent
+
+- Tool-calling architecture
+
+- Multi-step task execution
+
+- Project-aware workflows
+
+- Plan Mode and Build Mode
+
+- Error investigation
+
+- Task completion summaries
+
+- File change and diff review
+
+- Checkpoints and recovery
+  
+    </td>
+  <td width="50%" valign="top">
+
+🎨 Desktop Workspace
+
+- CustomTkinter interface
+
+- Chat-based development
+
+- Markdown responses
+
+- Project navigation
+
+- Model selection
+
+- Provider configuration
+
+- Activity Details panel
+
+- Integrated Skills management
+  
+    </td>
+</tr>
+<tr>
+  <td width="50%" valign="top">
+
+🧩 Extension Architecture
+
+- Native Python Skills
+
+- Application Add-ons
+
+- MCP integration
+
+- Custom agent tools
+
+- Namespaced commands
+
+- Application lifecycle hooks
+
+- Supported UI extensions
+
+- Service integrations
+  
+    </td>
+  <td width="50%" valign="top">
+
+🖥️ Execution Environments
+
+- Local execution
+
+- Alpine Linux sandbox
+
+- SSH remote execution
+
+- Browser automation
+
+- Configurable permissions
+
+- Project workspace isolation
+
+- Multiple execution targets
+  
+    </td>
+</tr>
+
+</table>---
+
+🧠 Three Independent Extension Systems
+
+Agent Code 2.0 has three distinct extension systems.
+
+They are designed for different purposes and can coexist.
+
+System| Purpose| Extension type
+🧩 Skills| Extend what the AI agent can do| Native Python
+🔌 Add-ons| Extend Agent Code itself| Application extensions
+🌐 MCP| Connect external tools| MCP protocol
+
+🧩 1. Skills — Extend the AI Agent
+
+Skills give the AI agent additional tools and specialized capabilities.
+
+A Skill can provide:
+
+- Python functions.
+- AI-callable tool schemas.
+- Agent-specific commands.
+- Natural-language command aliases.
+- Skill metadata.
+- Help documentation.
+
+Example Skill structure
+
+skills/
+└── my_skill/
+    ├── skill.py
+    └── skill.md
+
+Example implementation
+
+SKILL = {
+    "name": "my_skill",
+    "description": "An example Agent Code skill."
+}
+
+
+def greet(name: str = "world") -> str:
+    return f"Hello, {name}!"
+
+
+TOOL_SCHEMAS = [
+    {
+        "type": "function",
+        "function": {
+            "name": "greet",
+            "description": "Return a greeting.",
+            "parameters": {
+                "type": "object",
+                "properties": {
+                    "name": {"type": "string"}
+                },
+                "required": []
+            }
+        }
+    }
+]
+
+TOOL_CALLABLES = {
+    "greet": greet
+}
+
+The native loader reads:
+
+Export| Purpose
+"SKILL"| Skill metadata
+"TOOL_SCHEMAS"| AI tool definitions
+"TOOL_CALLABLES"| Python implementations
+"COMMANDS"| Optional commands and aliases
+"HELP"| Optional help text
+
+Skills can be reloaded without restarting Agent Code. When an agent task is running, tool reloads are deferred until the task finishes.
+
+Bundled Skills
+
+Agent Code 2.0 includes:
+
+Skill| Description
+📁 "filesystem"| File operations within the workspace
+🌐 "public_api"| Search and use public APIs
+📧 "temp_mail"| Temporary inbox and supported email integrations
+
+Install bundled Skills:
+
+agent-code skills install
+
+List bundled Skills:
+
+agent-code skills list
+
+Developer documentation: "Build a Skill" (docs/skills.html)
+
+---
+
+🔌 2. Add-ons — Extend Agent Code Itself
+
+Skills extend the AI agent.
+
+Add-ons extend the application.
+
+This is one of the central architectural differences in Agent Code 2.0.
+
+Add-ons are designed to let developers customize supported parts of the application, rather than being limited to adding AI tools.
+
+Add-on capabilities
+
+- Custom application UI.
+- Sidebar contributions.
+- Additional application actions.
+- Workspace extensions.
+- Settings extensions.
+- Toolbar contributions.
+- Agent tools and commands.
+- Lifecycle hooks.
+- Supported service replacements.
+
+Example structure
+
+my-addon/
+├── addon.json
+└── addon.py
+
+Example manifest
+
+{
+    "name": "my-addon",
+    "version": "1.0.0",
+    "description": "An example Agent Code extension",
+    "author": "Developer"
+}
+
+Example Add-on
+
+def build_sidebar_button(parent, app):
+    import customtkinter as ctk
+
+    return ctk.CTkButton(
+        parent,
+        text="My Workspace",
+        command=lambda: app.log_line(
+            "Workspace opened",
+            "dim"
+        )
+    )
+
+
+def register(api):
+    api.register_ui(
+        "my-workspace",
+        build_sidebar_button,
+        location="sidebar",
+        label="My Workspace"
+    )
+
+    api.register_action(
+        "Sample action",
+        lambda: "Action completed"
+    )
+
+Supported UI locations
+
+Location| Purpose
+"sidebar"| Sidebar components
+"addons"| Add-ons page
+"settings"| Settings contributions
+"toolbar"| Toolbar components
+"workspace"| Workspace components
+
+Add-ons can also register named service implementations where the relevant Core integration exposes a service extension point.
+
+They use the supported Add-on API rather than unrestricted monkey-patching of application internals.
+
+Security: Add-ons are trusted Python code and execute with the user's operating-system permissions. They are not sandboxed.
+
+Developer documentation: "Add-on Development" (ADDON-DEVELOPMENT.md)
+
+---
+
+🌐 3. MCP — Connect External Tools
+
+Agent Code supports Model Context Protocol integration.
+
+MCP allows the AI agent to connect to external tool servers and use capabilities exposed by those servers.
+
+This provides another way to extend the agent without implementing every external integration as a native Skill.
+
+MCP is managed separately from native Skills and application Add-ons.
+
+Skills, Add-ons, and MCP are independent systems.
+
+---
+
+⚙️ Agent Modes and Permissions
+
+Agent Code separates the way an agent approaches a task from the permissions granted to its operations.
+
+📋 Plan Mode
+
+Plan Mode focuses on understanding a task before implementation.
+
+Useful for:
+
+- Analyzing an existing project.
+- Understanding requirements.
+- Identifying files that may need changes.
+- Preparing implementation steps.
+- Reviewing an approach before building.
+
+🛠️ Build Mode
+
+Build Mode focuses on implementing the requested work.
+
+The agent can use its available tools to:
+
+- Create files.
+- Modify source code.
+- Execute supported commands.
+- Investigate errors.
+- Implement planned changes.
+- Review task results.
+
+⚡ Core Mode
+
+Core Mode provides a lightweight agent configuration with a reduced tool surface.
+
+It focuses on essential capabilities such as file creation, shell execution, and CURL operations.
+
+Skills and Add-ons behavior follows the application's Core Mode policy.
+
+🔐 Permission behavior
+
+Agent Code provides configurable permission behavior.
+
+Mode| Purpose
+🛡️ Safe| Restrict potentially sensitive operations
+⚙️ Normal| Balanced execution permissions
+⚡ Auto| Automatically allow operations covered by the configured policy
+👁️ Read-only| Restrict operations that modify files or execute commands
+
+Automatic permission does not override remote-host restrictions, operating-system permissions, or unavailable execution environments.
+
+---
+
+🖥️ Execution Environments
+
+Agent Code separates tool capabilities from their execution environment.
+
+💻 Local Execution
+
+Run supported operations on the host computer.
+
+Useful for:
+
+- Local development tools.
+- Project files.
+- Installed software.
+- Local configuration.
+- Browser access.
+
+When the optional sandbox is disabled, shell commands execute on the host OS.
+
+📦 Alpine Linux Sandbox
+
+Agent Code supports an optional Podman-based Linux sandbox.
+
+The sandbox uses:
+
+Alpine Linux 3.22
+
+It provides:
+
+- Bash.
+- Core Linux utilities.
+- Git.
+- CURL.
+- Python and pip.
+- GCC and G++ through Alpine build packages.
+- Alpine package management.
+
+Install additional packages inside the container using:
+
+apk add --no-cache package-name
+
+The project workspace is mounted into the container at:
+
+/workspace
+
+The sandbox uses a separate container environment, dropped capabilities, and "no-new-privileges".
+
+However, the project workspace remains shared with the host. Containerization is not a guarantee against every form of malicious or destructive operation.
+
+🌐 SSH Remote Execution
+
+Connect Agent Code to remote development machines.
+
+The SSH integration supports:
+
+- Hostname or IP.
+- SSH username.
+- Custom port.
+- Private-key path.
+- SSH agent.
+- OpenSSH configuration.
+- Password authentication.
+
+Remote commands are subject to the configured permission policy.
+
+Unknown SSH host keys are not silently trusted. Verify the host fingerprint before accepting it.
+
+Install the optional SSH dependency if required:
+
+python -m pip install paramiko
+
+Only connect to machines you are authorized to administer.
+
+---
+
+🌍 Browser Automation
+
+Agent Code includes browser automation capabilities through its browser integration.
+
+Browser tools can support workflows involving:
+
+- Opening websites.
+- Navigating pages.
+- Interacting with web interfaces.
+- Inspecting page content.
+- Browser diagnostics.
+
+A supported Chromium-based browser must be installed on the machine where browser automation runs.
+
+Firefox and Chromium browser detection is included in the diagnostic workflow, but browser compatibility depends on the automation backend and installed browser components.
+
+If automatic browser detection fails, check the browser executable configuration.
+
+Run:
+
+agent-code doctor
+
+---
+
+🤖 AI Providers and Models
+
+Agent Code uses an OpenAI-compatible client architecture.
+
+You can configure supported providers through the application's API and model settings.
+
+Supported provider presets
+
+Provider| Endpoint
+OpenAI| "https://api.openai.com/v1"
+Google AI Studio| "https://generativelanguage.googleapis.com/v1beta/openai/"
+Anthropic| "https://api.anthropic.com/v1/"
+OpenRouter| Provider preset
+Custom| User-configured OpenAI-compatible endpoint
+
+Provider compatibility depends on supported API features, model identifiers, and request formats.
+
+The Anthropic compatibility endpoint uses the OpenAI SDK compatibility surface. Some native Claude API features may not be available through that interface.
+
+🔑 Bring Your Own Key
+
+Agent Code is designed around BYOK.
+
+You control:
+
+- Your AI provider.
+- Your API credentials.
+- Your selected model.
+- Your API usage.
+- Your provider spending.
+
+There is no mandatory recurring AI subscription built into the application.
+
+External providers may charge for their APIs or impose their own account requirements.
+
+🔒 Credential storage
+
+Provider credentials are stored through Python "keyring" in the operating system's credential store where supported.
+
+Examples include:
+
+- Windows Credential Manager.
+- macOS Keychain.
+- Linux Secret Service-compatible keyrings.
+
+The JSON configuration stores credential references rather than newly saved API-key values.
+
+If the operating-system credential store is unavailable, saving credentials fails rather than silently writing new secrets in plaintext.
+
+---
+
+🎨 Desktop Experience
+
+Agent Code uses Python and CustomTkinter for its desktop interface.
+
+The interface is organized around the development workspace.
+
+Workspace features
+
+- Chat-based interaction.
+- Project navigation.
+- Skills management.
+- Add-ons management.
+- MCP configuration.
+- API and model settings.
+- Help and documentation.
+- Task status and activity information.
+
+Activity Details
+
+Agent Code separates the main conversation from internal task activity.
+
+Activity Details OFF
+
+The chat remains focused on user messages and assistant responses.
+
+During a task, the compact status is:
+
+"Working in project"
+
+Internal settings, model events, checkpoints, and detailed task events stay outside the main conversation.
+
+Activity Details ON
+
+The activity panel provides access to detailed task and tool events, including completion information and recovery details.
+
+Task stopping
+
+Agent Code supports cooperative stop controls and a three-click force-stop gesture.
+
+The force-stop behavior requests stronger cancellation of the agent's work and further AI/tool communication.
+
+It is not designed to close the browser application.
+
+Some blocking network calls or operations may only stop when their timeout or current operation returns.
+
+Recovery and review
+
+Task completion reports can include:
+
+- Changed files.
+- Text additions and deletions where safely measurable.
+- Binary file identification.
+- Diff review.
+- Recovery information.
+
+Review file changes and restore conflicts before applying recovery actions.
+
+---
+
+📥 Installation
+
+Requirements
+
+- Python 3.10 or newer.
+- Tkinter.
+- CustomTkinter.
+- Internet access for remote AI providers.
+- A compatible browser for browser automation.
+
+Optional features require their respective dependencies.
+
+Install from PyPI
+
+The recommended installation method:
+
+python -m pip install --upgrade pip
+python -m pip install agent-code
+
+Upgrade:
+
+python -m pip install --upgrade agent-code
+
+Ubuntu / Debian
+
+Install the required system components:
+
+sudo apt update
+
+sudo apt install \
+    python3 \
+    python3-venv \
+    python3-tk \
+    chromium \
+    podman
+
+Podman is optional unless you want the Linux sandbox.
+
+Create a virtual environment:
+
+python3 -m venv .venv
+
+source .venv/bin/activate
+
+python -m pip install --upgrade pip
+
+python -m pip install agent-code
+
+Windows
+
+Install Python 3.10 or newer with Tcl/Tk support.
+
+Install a compatible Chromium-based browser for browser automation.
+
+Then:
+
+python -m pip install agent-code
+
+macOS
+
+Install Python 3.10 or newer with Tcl/Tk support.
+
+Install a compatible Chromium-based browser.
+
+Then:
+
+python3 -m pip install agent-code
+
+Install from source
+
+git clone https://github.com/agent-code-dev/agent-code.git
+
+cd agent-code
+
+python -m venv .venv
+
+source .venv/bin/activate
+
+python -m pip install --upgrade pip
+
+python -m pip install .
+
+On Windows, activate the virtual environment using:
+
+.venv\Scripts\activate
+
+---
+
+🚀 Getting Started
+
+Launch the desktop application
+
+agent-code
+
+Explicitly launch the GUI
+
+agent-code gui
+
+Launch the terminal interface
+
+agent-code tui
+
+The TUI is useful for headless Linux environments and SSH sessions.
+
+On Linux, Agent Code can automatically select the TUI when no graphical display is available.
+
+Check your installation
+
+agent-code doctor
+
+Generate machine-readable diagnostics:
+
+agent-code doctor --json
+
+Save a diagnostics report:
+
+agent-code doctor --output report.json
+
+The diagnostics report does not include API keys.
+
+Get command help
+
+agent-code --help
+
+---
+
+⚙️ First-Time Configuration
+
+1. Launch Agent Code.
+2. Open Settings.
+3. Navigate to the API and model settings.
+4. Select a provider.
+5. Enter your API key.
+6. Select a supported model.
+7. Save the configuration.
+8. Send a small test prompt.
+
+Choose the project directory you want Agent Code to work in.
+
+Project workspace files and chat data are associated with the active project, while Skills and Add-ons use per-user locations.
+
+---
+
+📁 Configuration and Data Locations
+
+Agent Code uses platform-specific directories for user configuration and extensions.
+
+Data| Linux / macOS| Windows
+Configuration| "~/.config/agent-code/"| "%APPDATA%\agent-code\"
+Skills| "~/.local/share/agent-code/skills/"| "%LOCALAPPDATA%\AgentCode\skills\"
+Add-ons| "~/.local/share/agent-code/addons/"| "%LOCALAPPDATA%\AgentCode\addons\"
+Project workspace| "./workspace/"| "./workspace/"
+Project chats| "./chats/"| "./chats/"
+
+The Linux directories respect the corresponding XDG environment variables.
+
+Legacy project-local Skills can be migrated into the global Skills directory. Existing global Skills are not overwritten during migration.
+
+---
+
+🧰 Command Reference
+
+Application commands
+
+Command| Description
+"agent-code"| Launch the application
+"agent-code gui"| Explicitly launch the desktop GUI
+"agent-code tui"| Launch the terminal interface
+"agent-code doctor"| Check the local environment
+"agent-code doctor --json"| Print diagnostic information as JSON
+"agent-code doctor --output report.json"| Save a diagnostic report
+"agent-code --help"| Display available commands
+
+Skills commands
+
+Command| Description
+"agent-code skills list"| List bundled Skills
+"agent-code skills install"| Install bundled Skills into the user/project environment
+
+In-chat commands
+
+The agent also supports built-in and extension-provided commands.
+
+Examples include:
+
+Command| Purpose
+"/reset"| Reset conversation memory
+"/history"| Display recent messages
+"/skills"| Inspect loaded Skills and commands
+"/save"| Save the conversation
+"/open <file>"| Open a workspace file
+
+Skills can register additional commands.
+
+Add-on commands are namespaced, for example:
+
+/sample-addon:status
+
+---
+
+🔐 Security
+
+Agent Code is an AI development tool that can execute commands and load extensions. Its security depends on the selected execution environment and the code you choose to trust.
+
+Local shell
+
+When Podman sandboxing is disabled:
+
+- Commands execute on the host operating system.
+- Shell operators and known high-risk command patterns are restricted.
+- Workspace boundaries are enforced where detectable.
+- It is not a strong security boundary.
+
+Podman sandbox
+
+When enabled:
+
+- Shell and background commands run in Alpine Linux 3.22.
+- The project workspace is mounted into the container.
+- Container capabilities are dropped.
+- "no-new-privileges" is enabled.
+- Outbound networking depends on Podman and host network configuration.
+
+The sandbox runs as root inside the container, not as host root.
+
+The workspace remains shared with the host, so commands can still change or delete project files.
+
+Extension security
+
+Component| Security consideration
+Skills| Arbitrary Python code
+Add-ons| Trusted Python code with user permissions
+MCP| External tools may access exposed resources
+Local shell| Executes on the host when sandboxing is disabled
+SSH| Executes commands on the connected remote machine
+
+Security recommendations
+
+- Review Skills before installing them.
+- Only install Add-ons from trusted developers.
+- Connect only trusted MCP servers.
+- Protect API credentials.
+- Keep SSH private keys secure.
+- Review commands before granting automatic permissions.
+- Inspect file diffs before accepting changes.
+- Treat web pages and repository content as untrusted input.
+
+---
+
+🛠️ Development
+
+Clone the repository:
+
+git clone https://github.com/agent-code-dev/agent-code.git
+
+cd agent-code
+
+Create a development environment:
+
+python -m venv .venv
+
+source .venv/bin/activate
+
+Install the project:
+
+python -m pip install --upgrade pip
+
+python -m pip install -e .
+
+Run tests:
+
+python -m pytest -q
+
+Build distribution packages:
+
+python -m pip install build
+
+python -m build
+
+The project uses a "src/"-based Python package structure.
+
+---
+
+📚 Documentation
+
+The repository includes offline-friendly HTML documentation.
+
+Documentation| Description
+📖 "User Guide" (docs/index.html)| Installation, configuration, workflows, permissions, browser, SSH and troubleshooting
+🧩 "Skills Guide" (docs/skills.html)| Skill structure, tool schemas, commands and development
+🔌 "Add-on Guide" (docs/addons.html)| Application extension API, UI contributions, actions, hooks and services
+📋 "Release Notes" (RELEASE-NOTES-2.0.0.md)| Agent Code 2.0.0 changes
+🖥️ "Server Documentation" (server/README.md)| Separate Skills registry backend
+
+---
+
+🗂️ Project Structure
+
+Agent-code/
+│
+├── src/
+│   └── agent_code/
+│       ├── __init__.py
+│       ├── __main__.py
+│       ├── agent.py
+│       ├── loader.py
+│       ├── addons.py
+│       ├── advanced_features.py
+│       ├── mcp_manager.py
+│       ├── paths.py
+│       ├── projects.py
+│       ├── skill_manager.py
+│       ├── doctor.py
+│       ├── cli.py
+│       ├── tui.py
+│       │
+│       └── data/
+│           ├── icons/
+│           └── skills/
+│
+├── docs/
+│   ├── index.html
+│   ├── skills.html
+│   └── addons.html
+│
+├── server/
+│   ├── skill-server.py
+│   ├── README.md
+│   └── LICENSE
+│
+├── tests/
+│
+├── ADDON-DEVELOPMENT.md
+├── RELEASE-NOTES-2.0.0.md
+├── README.md
+├── LICENSE
+├── MANIFEST.in
+└── pyproject.toml
+
+---
+
+🗺️ Roadmap
+
+Areas of continued development:
+
+- 🚀 Improve agent reliability and task execution.
+- 🧠 Reduce unnecessary AI context and token consumption.
+- 🧩 Expand the Skills ecosystem.
+- 🔌 Expand supported Add-on extension points.
+- 🌐 Improve MCP integration.
+- 📦 Improve sandbox reliability.
+- 🖥️ Improve cross-platform compatibility.
+- 🔐 Strengthen extension security.
+- 🎨 Improve the desktop experience.
+- ⚙️ Improve provider compatibility and configuration.
+
+---
+
+🤝 Contributing
+
+Contributions are welcome!
+
+You can help Agent Code by:
+
+- Reporting bugs.
+- Suggesting features.
+- Improving documentation.
+- Creating Skills.
+- Developing Add-ons.
+- Testing compatibility.
+- Reviewing architecture.
+- Submitting pull requests.
+
+For significant changes, open an issue before starting implementation so the proposed direction can be discussed.
+
+Repository: "agent-code-dev/agent-code" (https://github.com/agent-code-dev/agent-code)
+
+---
+
+📜 License
+
+Agent Code uses separate licenses for its application and standalone server.
+
+Component| License
+Agent Code application| MIT
+Standalone registry backend ("server/")| AGPL-3.0-only
+
+The server's AGPL license does not replace the application's MIT license.
+
+See:
+
+- "Application LICENSE" (LICENSE)
+- "Server LICENSE" (server/LICENSE)
+
+---
+
+<p align="center">
+  <img src="src/agent_code/data/icons/agent-code.png" width="65" alt="Agent Code">
+</p><h2 align="center">⚡ Agent Code 2.1</h2><p align="center">
+  <b>Your AI coding agent. Your tools. Your environment.</b>
+  <br><br>
+  <a href="https://github.com/agent-code-dev/agent-code">GitHub</a>
+  ·
+  <a href="https://pypi.org/project/agent-code/">PyPI</a>
+  ·
+  <a href="https://agent-code-dev.github.io">Website</a>
+</p><p align="center">
+  <sub>Built with Python · Designed for extensibility</sub>
+</p>