SShortSingh.
Back to feed

Developer Splits AI Docker Sandbox Into Three Repos After Using It to Ship iOS App

0
·1 views

A developer has publicly released a three-way split of their AI coding sandbox project, originally published in February as a single repository called ai-sandbox-dkmcp. The original repo combined a devcontainer template, a host-side MCP server, and an in-container tool discovery server into one codebase. Before making the split public, the developer privately restructured it into three separate repos — AI Sandbox, HostMCP, and SandboxMCP — and used that setup to build and ship a real iOS app called とけるよ. Building the app inside the sandbox exposed practical limitations, including the need for per-tool timeout declarations and Xcode-specific host tools, which led to meaningful improvements beyond what demo usage alone would have surfaced. Since going public, the individual repos have continued to receive updates, including a formal installer for HostMCP, host-path masking to prevent username leakage, and nested-git-repo detection in SandboxMCP.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Log in to join the discussion and vote.

Log in

Related stories

0
ProgrammingDEV Community ·

GQEngine: Lua-based 2D game engine built for simplicity and education

GQEngine is an experimental, open-source 2D game engine currently in early planning and development, created by Game Quest Studio. The project is written entirely in Lua and built around a modular plugin system, requiring no direct use of C or C++. It aims to include core game development features such as a state machine, event manager, camera, and scene system, all powered by LuaJIT and lua-sdl2. The engine is designed to simplify game development education and content production for the Game Quest Studio platform. The project is publicly available on GitHub, where the developer welcomes suggestions and community participation.

0
ProgrammingHacker News ·

Anthropic's Claude AI Found Capable of Identifying Cryptographic Weaknesses

Anthropic has published research exploring how its Claude AI model can be used to discover cryptographic vulnerabilities. The study suggests that large language models may have practical applications in the field of cryptographic security analysis. The research was shared on Anthropic's official website and drew attention from the Hacker News community. The findings raise important questions about both the offensive and defensive potential of AI in cybersecurity. Anthropic's work highlights a growing interest in leveraging AI tools to identify weaknesses in security systems before malicious actors can exploit them.

0
ProgrammingHacker News ·

OpenAI Explores Role of Agentic AI in Scientific Computing

OpenAI has published a piece examining how agentic AI systems are beginning to influence scientific computing workflows. The article explores the intersection of autonomous AI agents and computational science, suggesting a shift in how research tasks may be approached. Agentic AI refers to systems capable of taking multi-step actions toward goals with minimal human intervention. The discussion reflects growing industry interest in applying large language models and AI agents to complex scientific problem-solving.

Developer Splits AI Docker Sandbox Into Three Repos After Using It to Ship iOS App · ShortSingh