Aicoreutils
Let agents run sandboxed, JSON-first Unix-style file and shell operations through 114 coreutils commands with dry-run guardrails.
Overview
AI Coreutils is a Build-phase MCP server that exposes 114 JSON-first coreutils commands with sandbox and dry-run for AI coding agents.
What is this MCP server?
- 114 JSON-first Coreutils-style commands exposed for programmatic agent use
- Sandbox and dry-run modes to reduce destructive mistakes during agent-driven edits
- Dual surface: CLI plus MCP (PyPI package aicoreutils, manifest version 0.3.6)
- stdio transport via PyPI install—repository caseSHY/AI-CLI on GitHub
- Suited to repo hygiene, batch file transforms, and scripted agent workflows on a solo dev machine
- 114 commands documented in registry description
- Manifest version 0.3.6; PyPI package identifier aicoreutils
- Transport: stdio
Community signal: 1 GitHub stars.
What problem does it solve?
Agents that run raw shell commands are brittle and risky when solo builders need structured, previewable file and system operations.
Who is it for?
Power users automating repo maintenance and local DevOps with agents who want guardrailed coreutils instead of ad-hoc bash.
Skip if: Beginners who only need IDE formatting, or teams seeking hosted cloud infra MCP rather than on-machine CLI tools.
What do I get? / Deliverables
After installing aicoreutils and registering the stdio MCP server, agents execute cataloged commands with JSON output, sandbox, and dry-run controls.
- MCP-accessible catalog of 114 JSON-first utility commands
- Sandboxed and dry-run pathways for agent-driven filesystem work
- Repeatable local automation without hand-written shell wrappers
Recommended MCP Servers
Journey fit
Agent-safe CLI tooling sits in Build because it extends what your coding agent can do on disk and in pipelines while you implement automation. Agent-tooling is the shelf for MCP servers that wrap command suites purpose-built for LLM agents, not generic SaaS APIs.
How it compares
Sandboxed agent-oriented coreutils MCP, not a cloud deploy server or a web-scraping browser automation skill.
Common Questions / FAQ
Who is AI Coreutils MCP for?
Solo builders and agent-heavy developers who want JSON-structured file and utility commands with sandbox and dry-run when Claude Code, Cursor, or Codex touches the filesystem.
When should I use AI Coreutils MCP?
Use it during Build when scripting agent-driven refactors, batch file tasks, or local automation—especially when you need dry-run before destructive changes.
How do I add AI Coreutils MCP to my agent?
Install the PyPI package aicoreutils (version aligned with the MCP manifest), configure the stdio MCP entry in your client, and ensure Python runtime requirements from the caseSHY/AI-CLI repository are met.