Back to directory

agency-agents (Chinese edition)

A Chinese-language collection of 277 ready-to-use AI agent personas covering roles such as CEO, CTO, engineering, design, marketing and finance, for use with Claude Code, Cursor, Copilot and other coding tools, with an orchestrator that runs several agents together.

agency-agents (Chinese edition)

This is a Chinese-language edition of the agency-agents project: a set of 277 plug-and-play AI expert personas covering company roles such as CEO, CTO, CMO, COO, CPO and CFO, plus engineering, design, marketing, product, games, security, GIS and finance, across 20 departments. Each persona defines its own character, professional process and deliverables rather than a generic prompt template.

Beyond a full translation of the upstream project, this edition adds 64 personas built for the Chinese market, covering platforms such as Xiaohongshu, Douyin, WeChat, Feishu and DingTalk, along with cross-border e-commerce, government affairs, medical compliance, industrial Qt interfaces and mechanical design.

The personas work with more than 20 AI coding tools, including Claude Code, Cursor and Copilot, and can be run individually or coordinated automatically through the accompanying agency-orchestrator, which assigns tasks to multiple agents as a DAG. A native desktop client is also available for macOS, Windows and Linux.

More in Agent skills

See all 581
  • Provenance gate

    github.com

    A deterministic authorization gateway that sits between an AI agent and its tools, tracing where each parameter of a side-effect call came from, and blocking the call or asking a human when the source does not check out.

  • A markdown framework of system rules that restates Isaac Asimov's Four Laws of Robotics as a set of prioritized loss calculations for large-language-model agents to follow, published as a discovery file for agents and crawlers to read.

  • Vitruvius

    github.com

    An engineering research agent that runs a discover, read, synthesize, verify and review loop over technical questions, reading sources directly rather than inferring from titles, and publishing its own scored verifier benchmark, weaknesses included.