agency-agents-zh: High-Fidelity AI Agent Personas

repository·main·Indexed 30 days ago

https://github.com/jnmetacode/agency-agents-zh

A collection of 267 plug-and-play AI agent personas designed for complex workflows, technical engineering, and system governance. Includes specialized integrations for Aider, Antigravity, Claude Code, Cursor, Gemini CLI, and GitHub Copilot, with support for MCP Memory integration to provide agents with persistent, cross-session memory.

Tokens
530.9K
Snippets
482
Records
1.6K
Agent score
97%

What's inside agency-agents-zh

  1. Overview of agency-agents-zh

    main

    agency-agents-zh is a collection of 268 plug-and-play AI expert roles covering 20 departments such as Engineering, Design, Marketing, Product, Games, Security, GIS, and Finance. Unlike generic prompt templates, each agent has a unique persona, professional workflow, and specific deliverables.

    Key features:

    • 268 Agents: Includes 215 translated from the upstream version and 53 original agents tailored for the Chinese market (e.g., Xiaohongshu, Douyin, WeChat, Feishu, DingTalk, and specialized industrial/medical domains).
    • Tool Support: Compatible with 18 AI programming tools including Claude Code, Cursor, and Copilot.
    • Expertise: Agents are designed with specific thinking patterns and workflows (e.g., a Security Engineer follows OWASP Top 10 standards).

    You can browse all experts online without installation at ao.aiolaola.com/experts.

  2. Overview of Agency Agents Project Scale

    main
    The Agency Agents project is a collection of 268 AI agent personas organized into 19 specialized departments. It includes 215 translated agents and 53 original agents created specifically for the Chinese market, supporting 18 different tools. The agents are categorized into departments such as Engineering, Design, Marketing, Sales, Finance, HR, Legal, Supply Chain, Product, Project Management, Testing, Support, Spatial Computing, Game Development, Academic, GIS, and Security.
  3. Use the Paid Media Auditor persona for comprehensive account audits

    main

    The Paid Media Auditor is a specialized persona designed to perform systematic, high-fidelity audits of advertising accounts across platforms like Google Ads, Microsoft Ads, and Meta. It evaluates over 200 checkpoints covering account structure, tracking, bidding, creative, audience, and competitive positioning.

    Use this persona to:

    • Conduct full audits before taking over existing accounts.
    • Perform quarterly health checks on managed accounts.
    • Diagnose root causes for performance declines.
    • Assess readiness before scaling budgets (e.g., 'Can this account handle 2x budget?').
    • Verify tracking before major campaign launches.
    • Perform compliance audits for regulated industries (Medical, Finance, Legal).
  4. Use the Historian persona for historical analysis and world-building

    main
    The Historian persona is a research-oriented agent designed for historical analysis, periodization, and material culture expertise. Use this persona to validate historical consistency in settings, enrich content with authentic period details (food, clothing, technology), and challenge historical myths or Eurocentric biases. It operates using historiographical methodologies like the Annales School, microhistory, and long-term structural analysis.
  5. Use the Spatial Data Scientist persona for advanced spatial analysis

    main

    The Spatial Data Scientist is an expert persona for performing advanced statistical modeling, spatial econometrics, clustering, and predictive analysis on geospatial data. Unlike a standard GIS Analyst who focuses on cartography, this persona focuses on detecting hidden patterns, quantifying uncertainty, and ensuring statistical rigor.

    Core Capabilities:

    • Spatial Pattern Detection: Identifying statistically significant clusters (Hot/Cold spot analysis), detecting spatial autocorrelation (Moran's I, Geary's C, Getis-Ord G), and point pattern analysis (Kernel Density Estimation, Nearest Neighbor).
    • Spatial Regression & Modeling: Implementing OLS, Spatial Lag, Spatial Error models, and Geographically Weighted Regression (GWR). It handles spatial autocorrelation in residuals to avoid violating independence assumptions.
    • Interpolation & Prediction: Using Kriging, cokriging, and regression kriging to predict values at unobserved locations.
    • Network & Flow Analysis: Origin-Destination (OD) flow analysis, network spatial statistics (Network K-function), and connectivity modeling.

    When to use this persona:

    • Use when you need to move beyond mapping to find statistically significant patterns.
    • Use when you need to model spatial relationships or predict outcomes based on geography.
    • Do NOT use for standard cartography (use GIS Analyst), image-based ML feature extraction (use GeoAI/ML Engineer), or data cleaning/ETL (use Spatial Data Engineer).
  6. Use the Product Manager persona for product lifecycle management

    main

    The Product Manager persona (Alex) is designed to manage the entire product lifecycle, from requirement discovery and strategic planning to roadmap development, stakeholder alignment, and GTM (Go-To-Market) execution. The persona focuses on bridging the gap between business goals, user needs, and technical constraints.

    Core Principles:

    • Problem-First Approach: Always identify the underlying user pain point or business goal before accepting a feature request.
    • Validation Before Construction: Treat all feature ideas as hypotheses; validate them using user interviews, behavioral data, support signals, or competitive pressure before committing resources.
    • Outcome over Output: Focus on measurable impact (e.g., retention, activation) rather than just shipping features.
    • Scope Management: Actively prevent scope creep by evaluating every change request against current sprint goals.
    • Clear Communication: Ensure alignment (understanding the 'why' and the role) even if total consensus is not reached.
  7. Use the Anti-Money Laundering (AML) Report Template

    main

    Use this template to generate quarterly AML work reports. The report should include:

    • Overview: Counts for large-amount transaction reports, STR submissions, KYC updates, and sanction hits.
    • Suspicious Transaction Analysis: A table detailing case IDs, transaction features, amounts, risk types, and disposal results.
    • Risk Trends: Descriptions of new laundering/fraud methods, high-risk industries, and geographic distributions.
    • Compliance Status: Results of regulatory inspections, rectification progress, and staff training completion rates.
    # [YYYY年Q季度] 反洗钱工作报告
    
    ## 概况
    - **大额交易报告数**:[N] 笔,涉及金额 ¥[金额]
    - **可疑交易报告(STR)**:[N] 份,已提交 [N] 份
    - **客户身份识别(KYC)**:新增 [N] 人,重新识别 [N] 人
    - **制裁名单命中**:[N] 次,其中确认 [N] 次,误报 [N] 次
    
    ## 可疑交易分析
    ### 典型案例
    | 案例编号 | 交易特征 | 涉及金额 | 风险类型 | 处置结果 |
    |----------|----------|----------|----------|----------|
    | STR-001  | 短期内频繁小额转账后大额汇出 | ¥850,000 | 疑似洗钱 | 已上报 |
  8. Use the Brand Guardian Agent for brand strategy and identity management

    main

    The Brand Guardian (品牌守护者) is a specialized agent designed for developing comprehensive brand identities, maintaining cross-channel consistency, and managing strategic brand evolution.

    Core Capabilities

    • Brand Foundation Development: Creating purpose, vision, mission, values, and personality frameworks.
    • Visual Identity Systems: Designing logo systems, color palettes (including accessibility/WCAG compliance), typography, and spacing systems.
    • Brand Voice & Messaging: Establishing tone of voice, messaging architecture, and writing guidelines.
    • Brand Protection: Implementing trademark strategies, compliance monitoring, and crisis management.

    Standard Workflow

    1. Discovery & Strategy: Analyze business needs, competition, and target audience.
    2. Foundation Development: Create the strategic framework and visual/verbal standards.
    3. System Creation: Design scalable assets like logo variants, color systems, and typography hierarchies.
    4. Implementation & Protection: Build asset libraries, compliance monitoring processes, and stakeholder training.
  9. Use the WeChat Mini Program Developer persona

    main
    The WeChat Mini Program Developer is a full-stack engineering expert specializing in the WeChat ecosystem. This persona is designed to handle the complete development lifecycle: from requirement analysis and architecture design (WXML/WXSS/WXS) to WeChat ecosystem integration (Login, WeChat Pay, Subscription Messages), Cloud Development (Cloud Functions, NoSQL Database, Cloud Storage), and performance optimization (Subpackaging, setData optimization, and bundle size control).
  10. Understand the NEXUS Operating Model and Deployment Modes

    main

    NEXUS is a deployment framework that synchronizes individual AI agents into a unified network. It uses a seven-stage pipeline to prevent architectural contradictions, redundant labor, and quality gaps during handoffs.

    Deployment Modes

    Choose a mode based on your project scope and timeline:

    • NEXUS-Full: For enterprise-level product launches covering the full lifecycle. (Timeline: 12-24 weeks, uses all agents).
    • NEXUS-Sprint: For feature development or MVP construction. (Timeline: 2-6 weeks, uses 15-25 agents).
    • NEXUS-Micro: For bug fixes, content activities, or single deliverables. (Timeline: 1-5 days, uses 5-10 agents).
  11. Use the Business Case Framework

    main

    Use this framework to structure formal investment proposals and capital decisions. A complete business case must include:

    1. Executive Summary: One-page overview including the specific decision required (Approve/Reject), required investment ($ amount and duration), expected returns (NPV, IRR, Payback Period), and the decision deadline.
    2. Opportunity: The problem/opportunity being addressed, strategic alignment, and the cost of inaction.
    3. Proposed Solution: Details of the proposal, why it was chosen over alternatives, and key assumptions.
    4. Financial Analysis: Breakdown of one-time and recurring costs, projected revenue/savings (Years 1-3), annual net cash flow, NPV (at X% discount rate), IRR, and Payback Period.
    5. Risk Assessment: Identification of key risks (Probability/Impact), mitigation plans, and sensitivity analysis (e.g., impact of a 20% deviation in key assumptions).
    6. Implementation: Timeline (milestones), required resources (people, capital, systems), dependencies, and decision gates (go/no-go points).
    7. Recommendation & Next Steps: The recommended decision and immediate actions following approval.
    商业论证结构
    执行摘要(1 页)
    所需决策:[具体、二元——批准或否决]
    所需投资:$[金额],历时 [周期]
    预期回报:$[NPV] / [IRR]% / [回收期]
    建议:[推进 / 不推进 / 有条件推进]
    决策截止时间:[日期]
    
    财务分析
    投资:$[一次性] + $[每年持续]
    收入/节省:$[第 1 年] / $[第 2 年] / $[第 3 年]
    按年净现金流:[表格]
    [X]% 折现率下的 NPV:$[金额]
    IRR:[%]
    回收期:[月数]
    
    风险评估
    关键风险 1:[描述] —— 概率:高/中/低 —— 影响:高/中/低
    缓释:[我们如何降低此风险]
    敏感性:[若关键假设偏差 20% 会怎样?]
  12. Use the Law Firm Client Intake Expert persona

    main

    The Law Firm Client Intake Expert (律所客户接案专家) is a specialized AI persona designed to manage the entire client intake lifecycle for law firms. It handles potential client qualification, case information gathering, consultation scheduling, and conflict of interest screening.

    Core Capabilities:

    • Initial Contact: Warm greetings and initial needs assessment.
    • Qualification: Screening by case type, jurisdiction, urgency, and fee model compatibility.
    • Conflict Screening: Identifying parties and checking against existing firm clients/matters.
    • Information Gathering: Collecting facts, timelines, documents, and prior legal actions.
    • Scheduling: Matching clients with appropriate lawyers and coordinating calendars.
    • Intake Summaries: Providing "lawyer-ready" briefs before consultations.
    • Follow-up: Managing no-shows, client nurturing, and referral routing.