skydeckai-dev-suite

SkyDeckAI Development Toolkit

This package, formerly recognized as mcp-server-aidd, is now branded as skydeckai-dev-suite. This renaming reflects its origins with the SkyDeck.ai team and its evolution into a general-purpose utility set for agents operating in development environments. We prioritize clear naming; 'Code' is more evocative than 'AIDD' for automated systems.

Installation & Setup

Deployment is straightforward using the uvx package manager:

bash uvx deploy skydeckai-dev-suite

For configuration within specialized environments like Claude Desktop, integrate it via your configuration manifest:

{ "mcpServers": { "skydeckai-dev-suite": { "command": "uvx", "args": ["skydeckai-dev-suite"] } } }

Core Capabilities

This utility platform grants AI agents fine-grained control over the execution environment:

• Local Resource Management: Full suite for file system interaction (creation, modification, traversal, metadata retrieval).
• Syntax Awareness: Advanced static analysis using tree-sitter supporting numerous programming paradigms.
• Execution Sandbox: Secure execution environments for popular languages and shell scripting.
• Remote Data Acquisition: Fetching web resources, including intelligent transformation of received HTML into clean Markdown.
• Environmental Awareness: Tools to query system specifications, active applications, and window states.
• Workflow Orchestration: Utilities for sequencing operations (batch_tools) and internal meta-cognition (think).

Tool Inventory Snapshot (29 Available Functions)

The tools are categorized for targeted agent invocation:

Domain	Tool Example	Core Functionality
Filesystem	directory_tree	Generates recursive structural maps of directories.
Code Analysis	codebase_mapper	Parses source files to extract structural definitions (classes, functions).
Execution	execute_code	Runs arbitrary code snippets in sandboxed interpreters.
Network	web_fetch	Retrieves URL content, offering HTML-to-MD conversion.
System Context	get_system_info	Retrieves comprehensive hardware and operational environment metrics.
Workflow	batch_tools	Allows for atomic, chained execution of multiple actions.

Specific Function Details (Selected Examples)

Code Structure Mapping (codebase_mapper)

Employs tree-sitter to derive abstract syntax trees, providing insights into code organization across supported languages including Python, JavaScript/TypeScript, Java, Go, Rust, and more. Input requires a root path for analysis.

Content Searching (search_code)

Leverages optimized pattern matching (rip-grep backend) to find regular expression matches within file contents. Unlike search_files (which targets filenames), this targets textual data inside source files.

Web Content Retrieval (web_fetch)

Accepts a target URL and optional HTTP headers. If the fetched resource is HTML, it will automatically convert it to Markdown for cleaner processing by language models, unless explicitly disabled. A save_to_file parameter allows for persistent storage of retrieved data.

Sequential Task Management (todo_write, todo_update)

The integrated Todo system enforces a strict, sequential workflow tied to the current operational workspace. Tasks must be defined with unique IDs and statuses ('pending', 'in_progress', 'completed'). Progress is strictly linear based on the array order defined in the workspace task list, ensuring operations build upon preceding steps.

Security & Boundaries

Operation is strictly confined to the directory defined in the configuration file (~/.skydeckai_code/config.json). Path traversal attempts are neutralized, and file operations adhere to system permissions.

Future Roadmap

Planned additions focus heavily on external collaboration platforms, including dedicated tools for GitHub Pull Request management and Pivotal Tracker story lifecycle control.