GitHub trending, early June 12, 2026 snapshot: Xiaomi's MiMo-Code jumps ahead

Fastest-growing this period

MiMo-Code is Xiaomi's OpenCode fork for terminal AI coding, persistent project memory, subagents, MiMo Auto, and OpenAI-compatible provider setup.

+13

apple/container Swift

A tool for creating and running Linux containers using lightweight virtual machines on a Mac. It is written in Swift, and optimized for Apple silicon.

+10

msitarzewski/agency-agents Shell

A complete AI agency at your fingertips - From frontend wizards to Reddit community ninjas, from whimsy injectors to reality checkers. Each agent is a specialized expert with personality, processes, and proven deliverables.

mattpocock/skills Shell

Skills for Real Engineers. Straight from my .claude directory.

NVIDIA/SkillSpector Python

Security scanner for AI agent skills. Detect vulnerabilities, malicious patterns, and security risks.

microsoft/markitdown Python

Python tool for converting files and office documents to Markdown.

bannedbook/fanqiang Kotlin

翻墙-科学上网

chopratejas/headroom Python

Compress tool outputs, logs, files, and RAG chunks before they reach the LLM. 60-95% fewer tokens, same answers. Library, proxy, MCP server.

affaan-m/ECC JavaScript

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

tashfeenahmed/freellmapi TypeScript

OpenAI-compatible proxy that stacks the free tiers of 16 LLM providers (~1.7B tokens/month) behind one /v1 endpoint — plus any custom OpenAI-compatible endpoint. Smart routing, automatic failover, encrypted keys. Personal experimentation only.

Our read

The early June 12 snapshot is small in raw deltas, but the ordering is clear: XiaomiMiMo/MiMo-Code leads at #1 (+13), two days after launch, ahead of apple/container at #2 (+10). MiMo-Code is another OpenCode-lineage terminal agent, but its hook is memory: checkpoints, project knowledge, task logs, and a hosted MiMo Auto path for first setup. That is a familiar June pattern with a new owner name attached.

The rest of the top ten keeps the same shape as the week: agent workbenches, skill systems, and safety tooling around them. agency-agents and mattpocock/skills sit together at #3 and #4, while NVIDIA’s SkillSpector appears at #5 as a security scanner for agent skills. markitdown, headroom, and ECC show the supporting layer still matters: ingestion, context compression, and harness discipline are trending beside the agents themselves.