Show HN: Aict – Unix coreutils that output XML/JSON, built for AI agents (github.com)

The new tool "aict" offers a reimagined set of Unix core utilities that output structured XML or JSON, optimized for AI agents. Unlike traditional Unix commands that return human-readable plaintext, aict directly provides structured data accessible without parsing, significantly reducing the complexity of interaction for AI agents. For instance, tasks that previously required multiple calls to identify file type or language can now be accomplished with a single call, increasing efficiency and reducing potential errors. This development is significant for the AI/ML community as it enhances the capabilities of AI agents in processing system commands, allowing for more straightforward integration with tools like Claude Code. By implementing structured outputs with clear attributes for each field—such as file names, sizes, and MIME types—developers can minimize the ambiguity and fragility associated with text parsing. Although output costs in tokens are higher than traditional tools, the time saved and reduction in complexity can streamline workflows and improve performance in AI-driven applications. With features like built-in MCP server functionality, aict represents a notable advancement in creating more agile and effective AI toolchains.