GitHub Trending: 144 AI employee roles (12 departments) open-sourced for free, each with unique personalities, workflows, and KPIs

📄Full Article· Automatically extracted by trafilaturaGemini 翻譯1593 words

An open-source project called agency-agents has gone viral on GitHub, accumulating over 84,000 stars. It features more than 144 AI "employee" personas across 12 departments, including engineering, design, and marketing. Each persona comes with defined personality traits, workflows, and measurable success metrics, supporting over a dozen AI development tools such as Claude Code and Cursor. (Previous coverage: Google assembles an elite team to catch up with Anthropic, even its own engineers are secretly using Claude Code) (Background: Google DeepMind and MIT jointly develop AI agent CoDaS: capable of autonomous scientific research, writing papers in just 8 hours) On GitHub, an open-source project called agency-agents uses this logic to describe over 144 AIs, going viral with over 84,000 stars and 13,000 forks. Each employee has a name, personality, and KPIs. This is not a new AI model or a complex orchestration framework. The core asset is a collection of Markdown files, each serving as an onboarding manual for an AI employee, defining who it is, how it works, and what it delivers. Installing the project requires only a single command, which automatically deploys the agent configuration files to the directories of supported AI development tools. It currently supports over a dozen mainstream AI code editors, including Claude Code, GitHub Copilot, Cursor, Aider, and Windsurf. Once installed, developers can call the corresponding AI directly by its role name, prompting it to respond with a specific professional demeanor. The entire database is divided into 12 departments: Engineering, Design, Paid Media, Sales, Marketing, Product, Project Management, QA, Customer Support, Spatial Computing, Game Development, and Academic Research. Each of the 144+ agents possesses specialized domain knowledge and a specific job focus. The Engineering department alone has 27 agents, covering frontend, backend, mobile, AI, DevOps, cybersecurity, smart contracts, and incident commanders. The Marketing department has 25, including platform-specific roles for TikTok, Reddit, WeChat, Xiaohongshu, and Douyin. The traditional way of using AI is prompt engineering: writing a description to tell the AI what to do. The architecture of agency-agents is more optimized. Each persona file contains four core sections: domain expertise, personality and communication style, deliverables, and success metrics. Simply put, it is like an HR job description. A frontend engineer agent doesn't just know HTML and CSS; it has its own preferred architectural philosophy. A paid media strategist agent doesn't just execute ad operations; it follows a fixed review process and performance evaluation framework. Note: The project is licensed under MIT, free for both personal and commercial use, with no restrictions on derivative works. The project also includes a complete case study titled "Nexus Spatial Discovery": 8 agents from different departments working in parallel on a single product exploration task without human intervention. The final output includes: market validation and competitive analysis, complete system architecture (including SQL schema), brand strategy and visual identity direction, go-to-market plan and pricing strategy, customer support and community building roadmap, user experience personas and journey maps, as well as a 35-week execution plan and 65 sprint tickets. Multi-agent means letting "multiple AIs work simultaneously in a division of labor." This is equivalent to a full cross-departmental product exploration sprint, which usually takes a team several weeks to complete, but in this case, it is handled in parallel by AI. Behind the 84,000 stars is a collective response from developers to a single question: I don't need a smarter AI assistant; I need a well-organized AI team. The basic unit of AI collaboration is evolving from "asking a question" to "assigning a project." When AI has roles, responsibilities, and performance standards, its working relationship with humans begins to resemble management rather than just usage. The next step might be learning how to be a good AI manager.

Data Status✓ Full text extractedRead Original (動區 BlockTempo)

🔍Historical Similar Events· Keyword + Asset Matching1 items

2026-04-23

"Fastest Ethereum L2" MegaETH sets TGE for 4/30, a quick look at tokenomics that prioritizes KPIs over timelines

Similarity 120%關鍵字 kpi同分類 zh

💡 Currently matching via keywords + symbols (MVP) · Will be upgraded to embedding semantic search later

Raw Information

ID:c543be87ad

Source:動區 BlockTempo

Published:2026-04-21 03:40:16

Category:zh_news · Export Category zh

Symbols:Unspecified

Community Votes:+0 / −0 · ⭐ 0 Important · 💬 0 Comments