Auto-sync: 2026-04-21 04:02

2026-04-21 04:02:47 +08:00
parent cb7c11e14f
commit ac524d1ff5
81 changed files with 3115 additions and 20 deletions
--- a/wiki/sources/agency-agents-examples.md
+++ b/wiki/sources/agency-agents-examples.md
@@ -0,0 +1,49 @@
+---
+title: "Examples (Agency Agents)"
+type: source
+tags: [agency-agents, multi-agent, examples]
+date: 2026-04-21
+---
+
+## Source File
+- [[raw/Agent/agency-agents/examples/README.md]]
+
+## Summary
+- 核心主题：Agency Agents 多智能体编排示例集
+- 问题域：展示多个专业智能体并行协作的实际效果
+- 方法/机制：8 个专业智能体同时部署，在产品发现任务中协同工作
+- 结论/价值：证明全量智能体编排可产生连贯、相互引用的计划
+
+## Key Claims
+- 多智能体并行执行可产生连贯、交叉引用的计划，无协调开销
+- 从"发现机会"到"完整蓝图"可在单次会话中完成
+- 智能体编排展示了对复杂任务的端到端处理能力
+
+## Key Quotes
+> "What does it actually look like when the full agency collaborates?" — 展示多智能体协作的核心问题
+
+## Key Concepts
+- [[Multi-Agent Team]]：多智能体团队，每个智能体有独立角色、人格和优化模型
+- [[Agents Orchestrator]]：智能体编排器，负责任务分配和流程协调
+- [[并行智能体执行]]：多个智能体同时工作于不同子任务
+
+## Key Entities
+- [[Product Trend Researcher]]：市场验证与竞争格局分析
+- [[Backend Architect]]：系统架构、数据模型、API 设计
+- [[Brand Guardian]]：品牌定位与视觉识别
+- [[Growth Hacker]]：GTM 策略与增长计划
+- [[Support Responder]]：客户支持运营蓝图
+- [[UX Researcher]]：用户画像与旅程地图
+- [[Project Shepherd]]：项目执行计划与风险管理
+- [[XR Interface Architect]]：空间界面架构规范
+
+## Connections
+- [[The Agency]] ← 定义了 ← [[多智能体系统可靠性]] 架构模式
+- [[Multi-Agent Team]] ← 实现于 ← [[nexus-spatial-discovery.md]] 示例
+
+## Contradictions
+- 未发现与现有内容的冲突
+
+## Notes
+- 本文档是示例索引，实际详细案例在 nexus-spatial-discovery.md
+- 8 个智能体并行工作的成功案例，展示了多智能体系统的实际应用价值
--- a/wiki/sources/antigravity-readme.md
+++ b/wiki/sources/antigravity-readme.md
@@ -0,0 +1,24 @@
+---
+title: "Antigravity Integration"
+type: source
+tags: [agency, integrations, antigravity]
+date: 2026-04-20
+source_file: raw/Agent/agency-agents/integrations/antigravity/README.md
+---
+
+## Summary
+Antigravity Integration README documents installing the Agency roster as Antigravity skills under `~/.gemini/antigravity/skills/`. Each agent is prefixed with `agency-` to avoid name conflicts. It covers install, activation by slug, regeneration steps, and the `SKILL.md` frontmatter format.
+
+## Key Claims
+- Agents install as Antigravity skills and are prefixed `agency-` to prevent conflicts.
+- Install via `./scripts/install.sh --tool antigravity`, which copies files to `~/.gemini/antigravity/skills/`.
+- Activate skills by their slug (e.g., `agency-frontend-developer`).
+
+## Key Quotes
+> "Installs the full Agency roster as Antigravity skills. Each agent is prefixed with `agency-` to avoid conflicts with existing skills." — Antigravity Integration README
+
+## Connections
+- [[The Agency: AI Specialists Ready to Transform Your Workflow]] — roster source
+
+## Contradictions
+- None detected with existing wiki content at time of ingest.
--- a/wiki/sources/cursor-readme.md
+++ b/wiki/sources/cursor-readme.md
@@ -0,0 +1,25 @@
+---
+title: "Cursor Integration"
+type: source
+tags: [agency, integrations, cursor]
+date: 2026-04-20
+source_file: raw/Agent/agency-agents/integrations/cursor/README.md
+---
+
+## Summary
+Cursor Integration README explains generating project-scoped Cursor `.mdc` rule files from The Agency roster. Rules live under `.cursor/rules/<agent-slug>.mdc` in a project and can be activated inline (e.g., `@frontend-developer`) or made always-on via frontmatter configuration.
+
+## Key Claims
+- The Agency can convert agents into Cursor `.mdc` rule files which are project-scoped.
+- Installers create `.cursor/rules/<agent-slug>.mdc` when run from a project root via `./scripts/install.sh --tool cursor`.
+- Rules can be invoked inline with `@agent-slug` or set alwaysApply in frontmatter.
+
+## Key Quotes
+> "Converts the full Agency roster into Cursor `.mdc` rule files. Rules are project-scoped — install them from your project root." — Cursor Integration README
+
+## Connections
+- [[The Agency: AI Specialists Ready to Transform Your Workflow]] — roster source
+- Cursor project-scoped rules pattern
+
+## Contradictions
+- None detected with existing wiki content at time of ingest.
--- a/wiki/sources/gemini-cli-readme.md
+++ b/wiki/sources/gemini-cli-readme.md
@@ -0,0 +1,24 @@
+---
+title: "Gemini CLI Integration"
+type: source
+tags: [agency, integrations, gemini]
+date: 2026-04-20
+source_file: raw/Agent/agency-agents/integrations/gemini-cli/README.md
+---
+
+## Summary
+Gemini CLI Integration README describes packaging the Agency agents as a Gemini CLI extension installed to `~/.gemini/extensions/agency-agents/`. It documents generating integration files, installing the extension, the extension file layout (including `skills/<agent>/SKILL.md`), how to activate skills by name, and how to regenerate files.
+
+## Key Claims
+- The Agency provides a Gemini CLI extension that installs under `~/.gemini/extensions/agency-agents/`.
+- Generate integration files with `./scripts/convert.sh --tool gemini-cli` and install with `./scripts/install.sh --tool gemini-cli`.
+- Skills are organized under `skills/<agent>/SKILL.md` and are invoked by name in the CLI.
+
+## Key Quotes
+> "Packages all 61 Agency agents as a Gemini CLI extension. The extension installs to `~/.gemini/extensions/agency-agents/." — Gemini CLI Integration README
+
+## Connections
+- [[The Agency: AI Specialists Ready to Transform Your Workflow]] — roster source
+
+## Contradictions
+- None detected with existing wiki content at time of ingest.
--- a/wiki/sources/github-copilot-readme.md
+++ b/wiki/sources/github-copilot-readme.md
@@ -0,0 +1,24 @@
+---
+title: "GitHub Copilot Integration"
+type: source
+tags: [agency, integrations, copilot]
+date: 2026-04-20
+source_file: raw/Agent/agency-agents/integrations/github-copilot/README.md
+---
+
+## Summary
+GitHub Copilot Integration README documents that The Agency agents are compatible with GitHub Copilot's `.md` + YAML frontmatter format and need no conversion. It includes install instructions (script or manual copy), activation examples, and references the main README for the full roster.
+
+## Key Claims
+- Agents work natively with GitHub Copilot using Markdown files with YAML frontmatter; no conversion required.
+- Installation via `./scripts/install.sh --tool copilot` or manual copying to `~/.github/agents/` or `~/.copilot/agents/`.
+- Agents are referenced by name within Copilot sessions to activate their behavior.
+
+## Key Quotes
+> "The Agency works with GitHub Copilot out of the box. No conversion needed — agents use the existing `.md` + YAML frontmatter format." — GitHub Copilot Integration README
+
+## Connections
+- [[The Agency: AI Specialists Ready to Transform Your Workflow]] — The roster source
+
+## Contradictions
+- None detected with existing wiki content at time of ingest.
--- a/wiki/sources/kimi-readme.md
+++ b/wiki/sources/kimi-readme.md
@@ -0,0 +1,24 @@
+---
+title: "Kimi Code CLI Integration"
+type: source
+tags: [agency, integrations, kimi]
+date: 2026-04-20
+source_file: raw/Agent/agency-agents/integrations/kimi/README.md
+---
+
+## Summary
+Kimi Code CLI Integration README explains converting Agency agents into Kimi agent directories containing `agent.yaml` and `system.md`, installing them to `~/.config/kimi/agents/`, usage examples (including `--agent-file` and `--work-dir`), agent directory structure, `agent.yaml` format, regeneration steps, and troubleshooting tips.
+
+## Key Claims
+- Agency agents are converted into Kimi agent specifications with `agent.yaml` and `system.md`.
+- Generate integration files with `./scripts/convert.sh --tool kimi` and install with `./scripts/install.sh --tool kimi`.
+- Agents are installed to `~/.config/kimi/agents/` and loaded via `kimi --agent-file <path>`.
+
+## Key Quotes
+> "Converts all Agency agents into Kimi Code CLI agent specifications. Each agent becomes a directory containing `agent.yaml` (agent spec) and `system.md` (system prompt)." — Kimi Code CLI Integration README
+
+## Connections
+- [[The Agency: AI Specialists Ready to Transform Your Workflow]] — roster source
+
+## Contradictions
+- None detected with existing wiki content at time of ingest.
--- a/wiki/sources/multi-agent-workflow-landing-page-sprint.md
+++ b/wiki/sources/multi-agent-workflow-landing-page-sprint.md
@@ -0,0 +1,50 @@
+---
+title: "Multi-Agent Workflow: Landing Page Sprint"
+type: source
+tags: []
+date: 2026-04-21
+---
+
+## Source File
+- [[raw/Agent/agency-agents/examples/workflow-landing-page.md]]
+
+## Summary
+- 核心主题：多智能体协作完成单日 landing page 开发 sprint
+- 问题域：如何在一天内完成高转化率 landing page 的生产发布
+- 方法/机制：4 个专业 Agent（Content Creator、UI Designer、Frontend Developer、Growth Hacker）分阶段并行协作
+- 结论/价值：通过并行启动、合并点、反馈循环和时间盒机制实现高效交付
+
+## Key Claims
+- 并行 kickoff：Copy 和 Design 可同时进行，因二者相互独立
+- 合并点：Frontend Developer 需要 Content Creator 和 UI Designer 的输出才能开始
+- 反馈循环：Growth Hacker 审查后，Frontend Developer 应用修改
+- 时间盒：每个步骤有明确时间限制以防止范围蔓延
+
+## Key Quotes
+> "Ship a conversion-optimized landing page in one day using 4 agents." — 工作流目标陈述
+> "Parallel kickoff: Copy and design happen at the same time since they're independent" — 并行机制说明
+> "Merge point: Frontend Developer needs both outputs before starting" — 合并依赖关系
+
+## Key Concepts
+- [[Multi-Agent Workflow]]：多智能体协作架构，每个 Agent 有独立角色、人格、优化的模型，通过共享内存+私有上下文实现协同
+- [[Parallel Kickoff]]：并行启动，多个工作流同时开始执行的机制
+- [[Merge Point]]：合并点，多个输入汇合后触发下一阶段的关键节点
+- [[Feedback Loop]]：反馈循环，审查后应用修改的迭代机制
+- [[Time-boxed]]：时间盒，单一任务的最大时长限制，防止无限蔓延
+
+## Key Entities
+- [[Content Creator Agent]]：The Agency 项目中的内容创作智能体，负责编写 landing page 文案
+- [[UI Designer Agent]]：The Agency 项目中的 UI 设计智能体，负责设计布局和组件规格
+- [[Frontend Developer Agent]]：The Agency 项目中的前端开发智能体，负责构建 landing page
+- [[Growth Hacker Agent]]：The Agency 项目中的增长黑客智能体，负责转化率优化
+- [[FlowSync]]：虚构的 API 集成平台产品，用于工作流演示
+
+## Connections
+- [[Content Creator Agent]] ← writes_copy ← [[Multi-Agent Workflow]]
+- [[UI Designer Agent]] ← designs ← [[Multi-Agent Workflow]]
+- [[Frontend Developer Agent]] ← builds ← [[Multi-Agent Workflow]]
+- [[Growth Hacker Agent]] ← optimizes ← [[Multi-Agent Workflow]]
+- [[Frontend Developer Agent]] ← depends_on ← [[Content Creator Agent]] + [[UI Designer Agent]]
+
+## Contradictions
+
--- a/wiki/sources/opencode-readme.md
+++ b/wiki/sources/opencode-readme.md
@@ -0,0 +1,24 @@
+---
+title: "OpenCode Integration"
+type: source
+tags: [agency, integrations, opencode]
+date: 2026-04-20
+source_file: raw/Agent/agency-agents/integrations/opencode/README.md
+---
+
+## Summary
+OpenCode Integration README explains generating `.opencode/agents/<slug>.md` project-scoped agent files. The converter maps named colors to hex codes, sets `mode: subagent` so agents are invoked with `@agent-name`, and supports installing project-scoped files or installing globally via `--path`.
+
+## Key Claims
+- OpenCode agents are `.md` files with YAML frontmatter stored in `.opencode/agents/` and use `mode: subagent`.
+- The installer creates `.opencode/agents/<slug>.md` when run from a project root; global installation is possible with `--path`.
+- Generated agent frontmatter includes `name`, `description`, `mode: subagent`, and `color` (hex codes produced from named colors).
+
+## Key Quotes
+> "OpenCode agents are `.md` files with YAML frontmatter stored in `.opencode/agents/`. The converter maps named colors to hex codes and adds `mode: subagent` so agents are invoked on-demand via `@agent-name` rather than cluttering the primary agent picker." — OpenCode Integration README
+
+## Connections
+- [[The Agency: AI Specialists Ready to Transform Your Workflow]] — roster source
+
+## Contradictions
+- None detected with existing wiki content at time of ingest.
--- a/wiki/sources/support-analytics-reporter.md
+++ b/wiki/sources/support-analytics-reporter.md
@@ -0,0 +1,45 @@
+---
+title: "Support Analytics Reporter"
+type: source
+tags: []
+date: 2026-04-21
+---
+
+## Source File
+- [[raw/Agent/agency-agents/support/support-analytics-reporter.md]]
+
+## Summary
+- 核心主题：数据分析与商业智能专家智能体定义
+- 问题域：数据驱动决策支持、商业洞察生成、仪表盘设计
+- 方法/机制：统计分析、RFM 客户分层、营销归因建模、预测模型、A/B 测试
+- 结论/价值：提供可量化的业务建议，实现 20%+ KPI 提升
+
+## Key Claims
+- 数据质量验证是分析的前提，必须在分析前完成数据准确性和完整性校验
+- 所有分析结论必须包含统计显著性测试和置信水平
+- 仪表盘设计需针对特定利益相关者需求和决策场景定制
+- 客户生命周期价值（CLV）计算是客户分析的核心指标
+
+## Key Quotes
+> "Be data-driven: Analysis of 50,000 customers shows 23% improvement in retention with 95% confidence"
+> "Focus on impact: This optimization could increase monthly revenue by $45,000 based on historical patterns"
+
+## Key Concepts
+- [[Data-Driven Decision Making]]：基于数据而非直觉的业务决策方法论
+- [[RFM Analysis]]：客户分层的经典方法，通过 Recency（最近购买）、Frequency（购买频率）、Monetary（消费金额）三个维度评估客户价值
+- [[Statistical Significance Testing]]：验证分析结果是否具有统计意义的假设检验方法
+- [[Marketing Attribution Modeling]]：多触点归因模型，将转化功劳分配给不同营销触点
+- [[Customer Lifetime Value]]：客户生命周期价值，衡量客户在整个关系周期内贡献的总收入
+- [[KPI Tracking]]：关键绩效指标监控，通过量化指标评估业务目标达成情况
+- [[Predictive Modeling]]：预测模型，基于历史数据预测未来趋势（流失、增长等）
+
+## Key Entities
+- [[The Agency]]：开源 AI 智能体集合项目，Analytics Reporter 是其销售与支持类别的智能体之一
+
+## Connections
+- [[Data Consolidation Agent]] ← supports ← [[Analytics Reporter]]
+- [[Sales Data Extraction Agent]] ← provides_data ← [[Analytics Reporter]]
+- [[Report Distribution Agent]] ← distributes_reports ← [[Analytics Reporter]]
+
+## Contradictions
+
--- a/wiki/sources/support-executive-summary-generator.md
+++ b/wiki/sources/support-executive-summary-generator.md
@@ -0,0 +1,65 @@
+---
+title: "Executive Summary Generator"
+type: source
+tags: []
+sources: []
+last_updated: 2026-04-21
+---
+
+## Source File
+- [[raw/Agent/agency-agents/support/support-executive-summary-generator.md]]
+
+## Summary
+- 核心主题：高级战略咨询风格的执行摘要生成智能体
+- 问题域：C-suite 决策者需要快速理解复杂业务信息
+- 方法/机制：McKinsey SCQA、BCG Pyramid Principle、Bain Action-Oriented Recommendation 三大框架
+- 结论/价值：将 325-475 字复杂业务输入转化为结构化执行摘要，支持 3 分钟决策
+
+## Key Claims
+- 咨询级 AI 系统应像资深战略顾问一样思考、结构和沟通
+- 执行摘要必须包含 ≥ 1 个量化或比较数据点
+- 量化影响应尽可能具体（Revenue/Cost/Market Share）
+- 建议必须包含 Owner + Timeline + Expected Result
+- 加速人类判断而非替代它
+
+## Key Quotes
+> "Think, structure, and communicate like a senior strategy consultant with Fortune 500 experience"
+> "Prioritize insight over information"
+> "Enable executives to grasp essence, evaluate impact, and decide next steps in under three minutes"
+
+## Key Concepts
+- [[McKinsey SCQA Framework]]：Situation-Complication-Question-Answer 结构化叙事框架
+- [[BCG Pyramid Principle]]：自上而下逻辑表达的的金字塔原则
+- [[Bain Action-Oriented Model]]：以行动为导向的建议模型，强调明确所有权和时间线
+- [[Executive Summary]]：325-475 字的决策者简报格式
+- [[Consulting Framework]]：管理咨询方法论，用于结构化复杂业务问题
+
+## Key Entities
+- [[The Agency]]：开源 AI 智能体集合项目，Executive Summary Generator 是其 Support 分类下的专业智能体
+
+## Connections
+- [[Data Consolidation Agent]] ← 应用场景 ← [[Executive Summary Generator]]（需要销售数据输入）
+- [[Report Distribution Agent]] ← 下游 ← [[Executive Summary Generator]]（生成摘要的分发）
+- [[Paid Media Auditor]] ← 类似框架 ← [[Executive Summary Generator]]（都需要量化发现）
+
+## Quality Standards
+- 325-475 words（≤ 500 max）
+- 每个关键发现包含 ≥ 1 个量化或比较数据点
+- 发现按业务影响排序
+- 建议包含 Owner + Timeline + Expected Result
+- 语气：Decisive、Factual、Outcome-driven
+
+## Workflow
+1. Intake and Analysis：审查业务内容，识别关键洞察和量化数据点
+2. Structure Development：应用金字塔原则，按业务影响优先级组织
+3. Executive Summary Generation：起草情境概述、关键发现（加粗战略含义）、量化业务影响、优先级建议
+4. Quality Assurance：验证字数范围、量化合规性、建议完整性
+
+## Output Format
+```
+## 1. SITUATION OVERVIEW [50-75 words]
+## 2. KEY FINDINGS [125-175 words]
+## 3. BUSINESS IMPACT [50-75 words]
+## 4. RECOMMENDATIONS [75-100 words]
+## 5. NEXT STEPS [25-50 words]
+```
--- a/wiki/sources/support-finance-tracker.md
+++ b/wiki/sources/support-finance-tracker.md
@@ -0,0 +1,38 @@
+---
+title: "Finance Tracker"
+type: source
+tags: [support, finance, agent]
+date: 2026-04-21
+---
+
+## Source File
+- [[raw/Agent/agency-agents/support/support-finance-tracker.md]]
+
+## Summary
+- 核心主题：企业财务健康维护、预算与差异分析、现金流预测与优化、投资评估
+- 问题域：现金流波动、预算超支、投资决策缺乏量化依据与审计合规
+- 方法/机制：年度/季度预算框架、滚动现金流预测、支付时点优化、NPV/IRR 投资分析、审计轨迹与合规校验
+- 结论/价值：通过制度化预算与预测、支付优化与投资评价，可提升现金流稳定性、提高投资回报率并保证审计合规性
+
+## Key Claims
+- 财务准确性优先：所有数据源与计算必须校验并记录假设与凭证
+- 预算与差异分析是发现业务偏差与触发纠正措施的首要机制
+- 滚动现金流预测与支付时点优化能显著缓解短期流动性风险
+- NPV/IRR/回收期等量化指标必须作为投资决策的基线，并结合风险评分做综合判断
+
+## Key Quotes
+> "Include financial compliance validation and audit trail documentation in all processes" — 强调审计链与合规验证为默认要求
+
+## Key Concepts
+- [[Cash Flow Management]]：滚动现金流预测、低现金预警与高现金投资机会识别
+- [[Investment Analysis]]：NPV、IRR、回收期、风险评分与投资建议逻辑
+- [[Budget Framework]]：年度预算、季度差异分析与部门预算汇总
+
+## Key Entities
+- [[Finance Tracker]]：定义的智能体人格（财务分析与控制）
+
+## Connections
+- [[Finance Tracker]] ← depends_on ← [[Support Analytics Reporter]]
+
+## Contradictions
+- 未发现与现有 wiki 内容的直接冲突
--- a/wiki/sources/support-infrastructure-maintainer.md
+++ b/wiki/sources/support-infrastructure-maintainer.md
@@ -0,0 +1,78 @@
+---
+title: "Infrastructure Maintainer"
+type: source
+tags: [agent, infrastructure, devops]
+date: 2026-04-21
+---
+
+## Source File
+- [[raw/Agent/agency-agents/support/support-infrastructure-maintainer.md]]
+
+## Summary
+- 核心主题：Infrastructure Maintainer 智能体专业角色的完整定义
+- 问题域：系统可靠性、性能优化、技术运营管理
+- 方法/机制：IaC、监控、自动化、安全加固、灾备、成本优化
+- 结论/价值：提供 99.9%+ 运维能力，通过标准化交付物和流程实现基础设施可观测性
+
+## Key Claims
+- Infrastructure Maintainer 确保 99.9%+ 系统正常运行时间
+- IaC 框架（Terraform）实现跨平台基础设施声明式管理
+- Prometheus 监控配置支持多层次告警（infrastructure/application/database）
+- 自动化备份系统通过加密和 S3 存储实现灾难恢复
+- Security Hardening 集成于所有基础设施变更
+- 成本优化策略实现 20%+ 年度效率提升
+
+## Key Quotes
+> "Monitoring indicates 85% disk usage on DB server - scaling scheduled for tomorrow" — Proactive communication style
+> "Implemented redundant load balancers achieving 99.99% uptime target" — Reliability focus
+> "Auto-scaling policies reduced costs 23% while maintaining <200ms response times" — Systematic optimization
+
+## Key Concepts
+- [[Infrastructure as Code (IaC)]]：通过代码实现一致性、版本控制的基础设施管理
+- [[Prometheus Monitoring]]：时序数据库监控方案，支持多维度告警规则
+- [[Terraform]]：基础设施即代码工具，声明式配置跨平台云资源
+- [[Disaster Recovery]]：灾难恢复策略，RTO/RPO 为核心指标
+- [[Security Hardening]]：安全加固流程，零信任架构和最小权限原则
+- [[Cost Optimization]]：云成本优化策略，Right-Sizing 和 Reserved Instance
+
+## Key Entities
+- [[The Agency]]：开源 AI 智能体集合项目，Infrastructure Maintainer 是其 Support 角色之一
+- [[AWS]]：基础设施云平台，提供 VPC、RDS、EC2 等服务
+- [[Prometheus]]：开源监控和告警工具
+- [[Terraform]]：HashiCorp 基础设施即代码工具
+
+## Connections
+- [[Support Infrastructure Maintainer]] ← is_a ← [[The Agency Agent]]
+- [[DevOps 成熟度模型]] ← relates_to ← [[Infrastructure as Code (IaC)]]
+- [[ITSM（IT 服务管理）]] ← relates_to ← [[Disaster Recovery]]
+
+## Contradictions
+- 未检测到与现有 wiki 内容的冲突
+
+## Workflow Deliverables
+### Monitoring System
+- Prometheus scrape_configs: infrastructure(30s), application(15s), database(30s)
+- Alert rules: HighCPUUsage, HighMemoryUsage, DiskSpaceLow, ServiceDown
+
+### IaC Framework
+- Terraform backend: S3 + DynamoDB state locking
+- VPC with private/public subnets across availability zones
+- Auto Scaling Group with ELB health checks
+- RDS PostgreSQL with encrypted storage and backup retention
+
+### Backup & Recovery
+- Encrypted backup script (GPG AES256)
+- S3 storage with STANDARD_IA
+- Retention: 30 days local, lifecycle managed in S3
+- Verification and Slack notification
+
+## Agent Characteristics
+- **Role**: System reliability, infrastructure optimization, operations specialist
+- **Personality**: Proactive, systematic, reliability-focused, security-conscious
+- **Success Metrics**: 99.9%+ uptime, MTTR <4 hours, 20%+ cost efficiency, 70%+ automation reduction
+
+## Advanced Capabilities
+- Multi-cloud architecture design
+- Container orchestration (Kubernetes)
+- Zero-trust security architecture
+- Compliance automation (SOC2, ISO27001)
--- a/wiki/sources/support-legal-compliance-checker.md
+++ b/wiki/sources/support-legal-compliance-checker.md
@@ -0,0 +1,41 @@
+---
+title: "Legal Compliance Checker"
+type: source
+tags: [agent, legal, compliance]
+date: 2026-04-21
+---
+
+## Source File
+- [[raw/Agent/agency-agents/support/support-legal-compliance-checker.md]]
+
+## Summary
+The Agency 项目中的法律合规检查专家智能体，专注于多司法管辖区监管合规（GDPR、CCPA、HIPAA、SOX、PCI-DSS）、风险评估、政策开发与合规监控，确保业务运营符合相关法律法规。
+
+## Key Claims
+- 合规优先方法：验证监管要求后再实施任何业务变更，文档化所有合规决策
+- 风险评估整合：对所有新业务举措和功能开发进行法律风险评估
+- 多司法管辖区合规验证：默认要求包含多辖区合规验证和审计追踪文档
+
+## Key Quotes
+> "GDPR Article 17 requires data deletion within 30 days of valid erasure request" — 精确援引法规条款
+> "Non-compliance with CCPA could result in penalties up to $7,500 per violation" — 风险量化
+> "Implemented consent management system achieving 95% compliance with user rights requirements" — 可度量成果
+
+## Key Concepts
+- [[GDPR Compliance Framework]]：欧盟通用数据保护条例合规框架
+- [[CCPA Compliance]]：加州消费者隐私法案合规
+- [[Privacy Policy Generator]]：隐私政策生成器 Python 类
+- [[Contract Review System]]：合同审查自动化系统
+- [[Regulatory Compliance Assessment Report]]：监管合规评估报告模板
+- [[Multi-Jurisdictional Compliance]]：多司法管辖区合规管理
+
+## Key Entities
+- [[The Agency]]：开源 AI 智能体集合项目，本智能体所属项目
+- [[Healthcare Marketing Compliance Specialist]]：同项目医疗健康合规智能体
+
+## Connections
+- [[The Agency]] ← provides ← [[Legal Compliance Checker]]
+- [[Healthcare Marketing Compliance Specialist]] ← related_to ← [[Legal Compliance Checker]]（共享合规方法论）
+
+## Contradictions
+- 本页面为 Agent 定义文档，与实际业务合规实施可能存在理论-实践差距
--- a/wiki/sources/support-support-responder.md
+++ b/wiki/sources/support-support-responder.md
@@ -0,0 +1,48 @@
+---
+title: "Support Responder"
+type: source
+tags: [agent, customer-service, the-agency]
+date: 2026-04-21
+---
+
+## Source File
+- [[raw/Agent/agency-agents/support/support-support-responder.md]]
+
+## Summary
+- 核心主题：The Agency 项目中的客户支持专家智能体，提供全渠道客户服务、问题解决和用户体验优化
+- 问题域：客户服务交付、质量保证、知识管理、主动式客户成功
+- 方法/机制：多层级支持框架（Tier 1/2/3）、SLA 响应机制、知识库管理、分析仪表盘
+- 结论/价值：将每次支持交互转化为品牌体验，通过 85% 首次联系解决率和 4.5+ CSAT 分数驱动客户满意度和留存
+
+## Key Claims
+- 首次响应时间 SLA：2 小时内，85% 首次联系解决率
+- 全渠道支持：email、chat、phone、social media、in-app messaging
+- 客户满意度目标：4.5/5 以上
+- 首次联系解决率目标：80%+
+- SLA 合规率目标：95%+
+- 知识库贡献目标：减少 25%+ 未来工单量
+
+## Key Quotes
+> "I understand how frustrating this must be - let me help you resolve this quickly" — 共情沟通风格
+> "Here's exactly what I'll do to fix this issue, and here's how long it should take" — 解决方案导向
+> "To prevent this from happening again, I recommend these three steps" — 主动预防思维
+
+## Key Concepts
+- [[Multi-Channel Support Framework]]：跨 email、chat、phone、social media、in-app messaging 的统一客户支持框架
+- [[Support Tiers]]：Tier 1（通用）、Tier 2（技术）、Tier 3（专家）的分层支持架构
+- [[Customer Success]]：从被动响应到主动成功干预的客户生命周期管理
+- [[Knowledge Base Management]]：自助服务资源、文档优化和交互式故障排除系统
+- [[Support Analytics]]：响应时间、解决率、CSAT 等关键指标的计算与趋势分析
+
+## Key Entities
+- [[The Agency]]：开源 AI 智能体集合项目，本智能体属于其中
+- Support Analytics：客户支持分析仪表盘类，负责指标计算和趋势识别
+- Knowledge Base Manager：知识库管理类，负责文章创建和优化
+- Tier 1/2/3 Support Teams：分层支持团队，各有明确的升级标准
+
+## Connections
+- [[Support Legal Compliance Checker]] ← shares_framework ← [[Support Responder]]（共享合规框架）
+- [[Product Feedback Synthesizer]] ← depends_on ← [[Support Responder]]（客户反馈数据输入）
+- [[Sales Discovery Coach]] ← parallel ← [[Support Responder]]（客户需求发现方法论）
+
+## Contradictions
--- a/wiki/sources/testing-accessibility-auditor.md
+++ b/wiki/sources/testing-accessibility-auditor.md
@@ -0,0 +1,73 @@
+---
+title: "Accessibility Auditor"
+type: source
+tags: [accessibility, testing, agent, the-agency]
+date: 2026-04-21
+---
+
+## Source File
+- [[raw/Agent/agency-agents/testing/testing-accessibility-auditor.md]]
+
+## Summary
+- 核心主题：The Agency 项目中的无障碍审计专家智能体，基于 WCAG 2.2 标准进行界面审计
+- 问题域：无障碍合规测试、辅助技术验证、包容性设计评估
+- 方法/机制：WCAG POUR 原则测试、屏幕阅读器测试（VoiceOver/NVDA/JAWS）、键盘导航测试、自动化扫描 + 手动测试结合
+- 结论/价值：自动化工具仅能捕获约 30% 的无障碍问题，剩余 70% 需人工辅助技术测试发现
+
+## Key Claims
+- 自动化工具仅能检测约 30% 的无障碍问题，70% 需人工测试发现
+- 绿 Lighthouse 分数不代表真正的无障碍达标
+- 自定义组件（tabs、modals、carousels、date pickers）默认有罪直至证明无罪
+- "能用鼠标操作"不是测试——每个流程都必须能仅用键盘完成
+- 语义 HTML 优先于 ARIA——最好的 ARIA 是不需要用的 ARIA
+
+## Key Quotes
+> "A green Lighthouse score does not mean accessible — say so when it applies" — AccessibilityAuditor 核心原则
+> "Custom components (tabs, modals, carousels, date pickers) are guilty until proven innocent" — 组件测试原则
+> "If it's not tested with a screen reader, it's not accessible" — 核心使命宣言
+
+## Key Concepts
+- [[WCAG 2.2]]：Web Content Accessibility Guidelines 2.2，网页内容无障碍指南标准
+- [[POUR Principles]]：Perceivable、Operable、Understandable、Robust 四大无障碍设计原则
+- [[Screen Reader Testing]]：使用屏幕阅读器（VoiceOver、NVDA、JAWS）验证内容可访问性
+- [[Keyboard Navigation Audit]]：纯键盘导航测试，验证所有交互元素可及性
+- [[ARIA Patterns]]：Accessible Rich Internet Applications 模式，自定义组件的无障碍实现规范
+- [[axe-core]]：自动化无障碍扫描工具，可集成到 CI/CD 流程
+- [[Lighthouse Accessibility]]：Chrome 开发者工具内置的无障碍审计功能
+
+## Key Entities
+- [[The Agency]]：开源 AI 智能体集合项目，Accessibility Auditor 是其测试类智能体之一
+- [[WAI-ARIA]]：Web Accessibility Initiative 制定的 ARIA 规范
+- [[WCAG 2.2]]：W3C 发布的网页内容无障碍指南 2.2 版本
+
+## Connections
+- [[Testing Reality Checker]] ← collaborates_with ← [[Accessibility Auditor]]
+- [[Evidence Collector]] ← provides_evidence_to ← [[Accessibility Auditor]]
+- [[Frontend Developer]] ← receives_audit_from ← [[Accessibility Auditor]]
+- [[UI Designer]] ← receives_feedback_from ← [[Accessibility Auditor]]
+
+## Contradictions
+- 暂无已知冲突
+
+## Testing Methodology
+### Automated Scanning
+- axe-core CLI: `npx @axe-core/cli http://localhost:8000 --tags wcag2a,wcag2aa,wcag22aa`
+- Lighthouse: `npx lighthouse http://localhost:8000 --only-categories=accessibility --output=json`
+
+### Manual Testing Protocol
+1. **Screen Reader Testing**: VoiceOver (macOS)、NVDA (Windows)、JAWS 各平台完整流程测试
+2. **Keyboard Navigation**: 纯键盘完成所有关键用户旅程
+3. **Visual Testing**: 200%/400% 缩放、高对比度模式、减少动画模式
+4. **Component Deep Dive**: 自定义组件 WAI-ARIA Authoring Practices 对标审查
+
+### Severity Classification
+- **Critical**: 完全阻断部分用户访问
+- **Serious**: 需要变通方案的主要障碍
+- **Moderate**: 造成困难但有变通方案
+- **Minor**: 降低可用性的烦扰问题
+
+## Audit Deliverables
+- Accessibility Audit Report（含问题详情、WCAG 条款引用、修复建议）
+- Screen Reader Testing Protocol（按组件的 PASS/FAIL 表格）
+- Keyboard Navigation Audit（全局导航 + 组件特定模式检查清单）
+- Remediation Priority（Immediate/Short-term/Ongoing 三级优先级）
--- a/wiki/sources/testing-api-tester.md
+++ b/wiki/sources/testing-api-tester.md
@@ -0,0 +1,24 @@
+---
+title: "API Tester (Agent)"
+type: source
+tags: [agent, testing, api]
+date: 2026-04-20
+source_file: raw/Agent/agency-agents/testing/testing-api-tester.md
+---
+
+## Summary
+API Tester is an Agency agent specialized in comprehensive API validation, performance testing, and security-focused QA. The source file defines its identity, mission, testing workflows, technical deliverables (including example test suites), and success metrics.
+
+## Key Claims
+- The agent enforces a security-first testing approach and performance excellence standards (95th percentile <200ms target, 10x load validation).
+- Provides example automated test suites (Playwright-based) and deliverable templates for testing reports.
+- Workflow covers discovery, strategy, implementation, and monitoring with integration into CI/CD.
+
+## Key Quotes
+> "You are **API Tester**, an expert API testing specialist who focuses on comprehensive API validation, performance testing, and quality assurance." — Agent frontmatter
+
+## Connections
+- [[Testing Practices]] — related concept for strategies and tooling
+
+## Contradictions
+- None detected with existing wiki content at time of ingest.
--- a/wiki/sources/testing-evidence-collector.md
+++ b/wiki/sources/testing-evidence-collector.md
@@ -0,0 +1,109 @@
+---
+title: "Evidence Collector"
+type: source
+tags: [testing, qa, evidence, agency]
+date: 2026-04-21
+---
+
+## Source File
+- [[raw/Agent/agency-agents/testing/testing-evidence-collector.md]]
+
+## Summary
+- 核心主题：QA 证据收集智能体 EvidenceQA 的角色定义与测试方法论
+- 问题域：AI Agent 开发中的质量保证流程，避免无证据的"幻想式"报告
+- 方法/机制：基于 Playwright 截图、可复现命令、事实检查的证据驱动 QA
+- 结论/价值：建立现实的质量评估标准，默认发现 3-5 个问题，要求视觉证据
+
+## Key Claims
+- 视觉证据是唯一真相：无法截图证明的功能视为不存在
+- 默认发现问题：首次实现总有 3-5+ 个问题，"零问题"是危险信号
+- 一切需证明：每个声明都需要截图证据支撑
+- 诚实质量评估：Basic/Good/Excellent 级别，不接受虚假的 A+ 评分
+
+## Key Quotes
+> "Screenshots Don't Lie" — 视觉证据是唯一真相
+> "Default to Finding Issues" — 首次实现总有 3-5+ 个问题
+> "Prove Everything" — 每个声明都需要截图证据
+
+## Key Concepts
+- [[证据驱动 QA（Evidence-Driven QA）]]：要求所有声明都有视觉证据支撑的 QA 方法论
+- [[幻想式报告（Fantasy Reporting）]]：无证据支撑的乐观声明，如"零问题"、"完美评分"
+- [[Reality Checker]]：通过实际命令和截图验证功能真实状态
+- [[Playwright 截图]]：自动化捕获界面截图作为 QA 证据
+
+## Key Entities
+- [[EvidenceQA]]：截图驱动的 QA 专家智能体，厌恶幻想式报告
+- [[Reality Checker]]：与 EvidenceQA 协同的质量检查智能体
+- [[Test Results Analyzer]]：测试结果分析与缺陷预测智能体
+
+## Connections
+- [[Reality Checker (Agent)]] ← complements ← [[Evidence Collector]]
+- [[Test Results Analyzer]] ← extends ← [[Evidence Collector]]
+- [[Evidence Collector]] ← part_of ← [[The Agency]]
+- [[The Agency]] ← contains ← [[Testing Agents]]
+
+## QA Report Template
+
+```markdown
+# QA Evidence-Based Report
+
+## 🔍 Reality Check Results
+**Commands Executed**: [List actual commands run]
+**Screenshot Evidence**: [List all screenshots reviewed]
+**Specification Quote**: "[Exact text from original spec]"
+
+## 📸 Visual Evidence Analysis
+**Comprehensive Playwright Screenshots**: responsive-desktop.png, responsive-tablet.png, responsive-mobile.png, dark-mode-*.png
+**What I Actually See**:
+- [Honest description of visual appearance]
+
+**Specification Compliance**:
+- ✅ Spec says: "[quote]" → Screenshot shows: "[matches]"
+- ❌ Spec says: "[quote]" → Screenshot shows: "[doesn't match]"
+
+## 📊 Issues Found (Minimum 3-5)
+1. **Issue**: [Specific problem]
+   **Evidence**: [Screenshot reference]
+   **Priority**: Critical/Medium/Low
+
+## 🎯 Honest Quality Assessment
+**Realistic Rating**: C+ / B- / B / B+ (NO A+ fantasies)
+**Design Level**: Basic / Good / Excellent
+**Production Readiness**: FAILED / NEEDS WORK / READY
+
+## 🔄 Required Next Steps
+**Status**: FAILED (default unless overwhelming evidence)
+**Re-test Required**: YES
+```
+
+## Testing Protocol
+
+### Accordion Testing
+- 对比展开前后的截图
+- 验证内容是否正确显示
+
+### Form Testing
+- 截图空表单、填写后表单
+- 验证提交、验证、错误提示
+
+### Mobile Responsive Testing
+- 1920x1080、768x1024、375x667 三种分辨率
+- 验证汉堡菜单、布局、配色
+
+### Dark Mode Testing
+- 验证深色模式切换功能
+- 检查截图中的 dark-mode-*.png
+
+## Automatic Fail Triggers
+
+### Fantasy Reporting Signs
+- 声称"零问题"
+- 完美评分（A+, 98/100）
+- 无证据的"豪华/高级"声明
+- 未测试就声称"生产就绪"
+
+### Visual Evidence Failures
+- 无法提供截图
+- 截图与声明不符
+- 截图中可见功能损坏
+- 基础样式被声称 为"豪华"
--- a/wiki/sources/testing-performance-benchmarker.md
+++ b/wiki/sources/testing-performance-benchmarker.md
@@ -0,0 +1,25 @@
+---
+title: "Performance Benchmarker (Agent)"
+type: source
+tags: [agent, testing, performance]
+date: 2026-04-20
+source_file: raw/Agent/agency-agents/testing/testing-performance-benchmarker.md
+---
+
+## Summary
+Performance Benchmarker is an Agency agent specializing in performance testing, benchmarking, and optimization across applications and infrastructure. The source includes example k6 test scripts, reporting generation, workflow steps, deliverable templates, and success metrics.
+
+## Key Claims
+- Focuses on load, stress, endurance, and scalability testing with statistical analysis and confidence intervals.
+- Provides k6 examples and report generation, Core Web Vitals targets, capacity planning guidance, and monitoring recommendations.
+- Emphasizes realistic testing, before/after validation, and continuous monitoring in CI/CD.
+
+## Key Quotes
+> "You are **Performance Benchmarker**, an expert performance testing and optimization specialist who measures, analyzes, and improves system performance across all applications and infrastructure." — Agent frontmatter
+
+## Connections
+- [[API Tester (Agent)]] — related testing agent for API validation
+- [[Workflow Optimizer (Agent)]] — related to process and systems optimization
+
+## Contradictions
+- None detected with existing wiki content at time of ingest.
--- a/wiki/sources/testing-reality-checker.md
+++ b/wiki/sources/testing-reality-checker.md
@@ -0,0 +1,24 @@
+---
+title: "Reality Checker (Agent)"
+type: source
+tags: [agent, testing, integration]
+date: 2026-04-20
+source_file: raw/Agent/agency-agents/testing/testing-reality-checker.md
+---
+
+## Summary
+Reality Checker is an Agency integration agent that enforces evidence-based production readiness: defaulting to "NEEDS WORK" unless overwhelming automated evidence (screenshots, test-results) supports "READY". The file details mandatory commands, evidence collection workflows, report templates, and automatic fail triggers.
+
+## Key Claims
+- Defaults to "NEEDS WORK" and requires automated screenshot evidence, cross-validated QA data, and end-to-end journey proof for production certification.
+- Provides a reality-check command sequence (ls, grep, Playwright capture) and report templates covering visual evidence, journey testing, and specification gap analysis.
+
+## Key Quotes
+> "Defaults to \"NEEDS WORK\" — requires overwhelming proof for production readiness." — Reality Checker frontmatter
+
+## Connections
+- [[Testing Practices]] — connects to QA and integration testing methodologies
+- [[API Tester (Agent)]] — related testing agent for API validation
+
+## Contradictions
+- None detected with existing wiki content at time of ingest.
--- a/wiki/sources/testing-test-results-analyzer.md
+++ b/wiki/sources/testing-test-results-analyzer.md
@@ -0,0 +1,26 @@
+---
+title: "Test Results Analyzer"
+type: source
+tags: [testing, quality, analytics]
+date: 2026-04-21
+source_file: raw/Agent/agency-agents/testing/testing-test-results-analyzer.md
+---
+
+## Summary
+Test Results Analyzer defines an agent persona specialized in analyzing test outputs, identifying failure patterns, building predictive defect models, assessing release readiness, and producing stakeholder-specific reports and dashboards.
+
+## Key Claims
+- Analyze test execution results across functional, performance, security, and integration testing using statistical methods
+- Identify failure patterns, perform root cause analysis, and surface high-impact quality risks
+- Build predictive models for defect-prone areas and provide confidence intervals for release readiness recommendations
+- Deliver executive summaries, technical reports, and automated alerts to stakeholders
+
+## Key Quotes
+> "Every test result must be analyzed for patterns and improvement opportunities." — core requirement
+
+## Connections
+- [[Model QA Specialist]] — related expertise in ML model auditing and statistical validation
+- [[Performance Benchmarker (Agent)]] — complementary agent for performance testing and benchmarking
+
+## Contradictions
+- No direct contradictions found with existing wiki content.
--- a/wiki/sources/testing-tool-evaluator.md
+++ b/wiki/sources/testing-tool-evaluator.md
@@ -0,0 +1,48 @@
+---
+title: "Testing Tool Evaluator"
+type: source
+tags: [agent, testing, tool-evaluation, the-agency]
+sources: []
+last_updated: 2026-04-21
+---
+
+## Source File
+- [[raw/Agent/agency-agents/testing/testing-tool-evaluator.md]]
+
+## Summary
+- 核心主题：Tool Evaluator Agent（工具评估专家智能体）定义与技术评估方法论
+- 问题域：企业工具选择、供应商评估、投资回报分析
+- 方法/机制：加权评分评估框架、TCO/ROI 定量分析、安全/集成/可扩展性评估、供应商管理
+- 结论/价值：90% 推荐准确率、85% 采用率、20% 成本降低、25% ROI 达成
+
+## Key Claims
+- Tool Evaluator Agent 通过加权评分体系实现客观工具评估
+- TCO（总拥有成本）分析包含隐藏成本和扩展费用
+- 评估流程必须包含安全、集成和成本分析三个默认要求
+- 评估方法论强调可重复性和透明性
+
+## Key Quotes
+> "Always test tools with real-world scenarios and actual user data" — 证据驱动评估原则
+> "Calculate total cost of ownership including hidden costs and scaling fees" — 成本分析核心要求
+
+## Key Concepts
+- [[Tool Evaluation Framework]]：加权评分工具评估框架，包含功能性、可用性、性能、安全性、集成、支持、成本七个维度
+- [[TCO（Total Cost of Ownership）]]：总拥有成本，包含许可、实施、培训、维护、集成、迁移、支持等全生命周期成本
+- [[ROI Analysis]]：投资回报率分析，通过多场景和敏感性分析评估工具价值
+- [[Vendor Management]]：供应商关系管理与合同优化策略
+
+## Key Entities
+- [[Tool Evaluator Agent]]：技术评估与战略工具采纳专家智能体，专注于 ROI 导向的工具选择
+
+## Connections
+- [[Test Results Analyzer]] ← complements ← [[Tool Evaluator Agent]]
+- [[Evidence Collector]] ← complements ← [[Tool Evaluator Agent]]
+- [[Performance Benchmarker]] ← complements ← [[Tool Evaluator Agent]]
+- [[Reality Checker]] ← enforces ← evidence-based evaluation
+
+## Contradictions
+- 与现有质量保障智能体存在功能重叠，但侧重点不同：
+  - Tool Evaluator 聚焦工具选择与供应商评估
+  - Test Results Analyzer 聚焦测试结果分析与缺陷预测
+  - Performance Benchmarker 聚焦性能测试与基准验证
+  - Reality Checker 聚焦生产就绪证据收集
--- a/wiki/sources/testing-workflow-optimizer.md
+++ b/wiki/sources/testing-workflow-optimizer.md
@@ -0,0 +1,25 @@
+---
+title: "Workflow Optimizer (Agent)"
+type: source
+tags: [agent, testing, workflow, optimization]
+date: 2026-04-20
+source_file: raw/Agent/agency-agents/testing/testing-workflow-optimizer.md
+---
+
+## Summary
+Workflow Optimizer is an Agency agent focused on process improvement, automation, and cross-functional workflow optimization. The source defines its mission, analysis frameworks, example Python optimization code, deliverable templates, and success metrics.
+
+## Key Claims
+- Emphasizes data-driven process improvement and human-centered design.
+- Provides code examples and frameworks for analyzing workflows, identifying opportunities, designing optimized flows, and calculating quantified impact.
+- Recommends phased implementation with quick wins, medium-term, and strategic initiatives.
+
+## Key Quotes
+> "You are **Workflow Optimizer**, an expert process improvement specialist who analyzes, optimizes, and automates workflows across all business functions." — Agent frontmatter
+
+## Connections
+- [[API Tester (Agent)]] — related testing and validation practices
+- [[Testing Practices]] — potential concept linking testing and workflow optimization
+
+## Contradictions
+- None detected with existing wiki content at time of ingest.
--- a/wiki/sources/windsurf-readme.md
+++ b/wiki/sources/windsurf-readme.md
@@ -0,0 +1,24 @@
+---
+title: "Windsurf Integration"
+type: source
+tags: [agency, integrations, windsurf]
+date: 2026-04-20
+source_file: raw/Agent/agency-agents/integrations/windsurf/README.md
+---
+
+## Summary
+Windsurf Integration README explains consolidating the full Agency roster into a single `.windsurfrules` file that is project-scoped. It documents installing from the project root, referencing agents by name in prompts, and regenerating rules with the convert script.
+
+## Key Claims
+- The Agency exports a single `.windsurfrules` file containing all rules; rules are project-scoped and installed from the project root.
+- Install via `/path/to/agency-agents/scripts/install.sh --tool windsurf` from your project root.
+- Activate agents by referencing their name in the prompt (e.g., "Use the Frontend Developer agent...").
+
+## Key Quotes
+> "The full Agency roster is consolidated into a single `.windsurfrules` file. Rules are project-scoped — install them from your project root." — Windsurf Integration README
+
+## Connections
+- [[The Agency: AI Specialists Ready to Transform Your Workflow]] — roster source
+
+## Contradictions
+- None detected with existing wiki content at time of ingest.
--- a/wiki/sources/workflow-book-chapter.md
+++ b/wiki/sources/workflow-book-chapter.md
@@ -0,0 +1,45 @@
+---
+id: workflow-book-chapter
+title: "Workflow Example: Book Chapter Development"
+type: source
+tags: [workflow, agent, writing]
+sources: [raw/Agent/agency-agents/examples/workflow-book-chapter.md]
+last_updated: 2026-04-21
+---
+
+## Source File
+- [[raw/Agent/agency-agents/examples/workflow-book-chapter.md]]
+
+## Summary
+- 核心主题：AI Agent 将粗糙素材转化为战略性的第一人称书籍章节草稿的工作流
+- 问题域：作者有语音笔记、片段或战略笔记，但缺乏干净的章节草稿
+- 方法/机制：Book Co-Author Agent 通过版本化草稿、编辑注释和明确修订问题实现结构化产出
+- 结论/价值：产出强化品类定位、保留作者声音、暴露开放编辑决策的章节
+
+## Key Claims
+- Book Co-Author Agent 将源素材转化为带编辑注释和下一步问题的版本化章节草稿
+- 工作流目标不是通用代笔，而是产生强化品类定位的章节
+- 草稿必须保持第一人称声音
+- 章节必须有一个清晰的承诺和内部逻辑
+- 声明必须与源素材关联或标记为假设
+- 移除通用激励语言
+- 输出以明确修订问题结束，而非模糊交接
+
+## Key Quotes
+> "The goal is not generic ghostwriting. The goal is to produce a chapter that strengthens category positioning, preserves the author's voice, and exposes open editorial decisions clearly."
+
+## Key Concepts
+- [[Book Co-Author Agent]]：将源素材转化为版本化章节草稿的 AI Agent
+- [[Versioned Draft]]：通过迭代循环产生的带版本号的章节草稿
+- [[Editorial Notes]]：对假设和证据缺口的明确标注
+- [[Feedback Loop]]：Agent 提供的结构化修订请求机制
+
+## Key Entities
+- [[The Agency]]：开源 AI 智能体集合项目，提供各类专业化 Agent
+
+## Connections
+- [[Book Co-Author Agent]] ← is_example_of ← [[The Agency]]
+- [[Versioned Draft]] ← depends_on ← [[Feedback Loop]]
+- [[Editorial Notes]] ← supports ← [[Book Co-Author Agent]]
+
+## Contradictions
--- a/wiki/sources/workflow-startup-mvp.md
+++ b/wiki/sources/workflow-startup-mvp.md
@@ -0,0 +1,62 @@
+---
+title: "Multi-Agent Workflow: Startup MVP"
+type: source
+tags: []
+date: 2026-04-21
+---
+
+## Source File
+- [[raw/Agent/agency-agents/examples/workflow-startup-mvp.md]]
+
+## Summary
+- 核心主题：多智能体协作从创意到 MVP 交付的完整工作流程
+- 问题域：4 周 MVP 开发项目，如何协调多个专业智能体顺序与并行工作
+- 方法/机制：7 个专业角色（Sprint Prioritizer、UX Researcher、Backend Architect、Frontend Developer、Rapid Prototyper、Growth Hacker、Reality Checker），通过顺序交接、并行工作、质量门控、上下文传递实现协作
+- 结论/价值：证明多智能体团队可以高效完成端到端产品开发
+
+## Key Claims
+- Sprint Prioritizer 将项目分解为 4 周冲刺计划，每周期有明确交付物和验收标准
+- UX Researcher 在 Week 1 并行运行，快速完成竞品分析和差异化定位
+- Backend Architect 基于 Sprint Prioritizer 和 UX Researcher 输出设计 API 和数据模型
+- Frontend Developer 与 Rapid Prototyper 在 Week 2 并行构建核心功能
+- Reality Checker 在 Week 2 中点和 Week 4 _launch 前设置两个质量门控
+- Growth Hacker 在 Week 3 开始启动策略规划，与 Frontend Developer 并行工作
+- Sequential Handoffs 确保每个智能体输出成为下一个智能体输入
+- Context Passing 要求粘贴完整智能体输出而非摘要
+
+## Key Quotes
+> "Copy-paste agent outputs between steps — don't summarize, use the full output" — 上下文传递原则
+> "If a Reality Checker flags an issue, loop back to the relevant specialist to fix it" — 质量门控反馈机制
+
+## Key Concepts
+- [[Sequential Handoffs]]：顺序交接，每个智能体输出直接作为下一智能体输入
+- [[Parallel Work]]：并行工作，Week 1 中 Sprint Prioritizer 和 UX Researcher 可同时运行
+- [[Quality Gates]]：质量门控，Reality Checker 在关键节点进行生产就绪评估
+- [[Context Passing]]：上下文传递，始终粘贴完整智能体输出而非摘要
+- [[Multi-Agent Team]]：多智能体团队，多角色协作架构
+- [[Sprint Planning]]：冲刺规划，4 周 MVP 交付的阶段性计划
+
+## Key Entities
+- [[Sprint Prioritizer]]：冲刺优先级智能体，负责将项目分解为 4 周冲刺计划
+- [[UX Researcher]]：用户体验研究智能体，负责竞品分析和用户访谈
+- [[Backend Architect]]：后端架构智能体，负责 API 设计、数据库 Schema 和实时架构
+- [[Frontend Developer]]：前端开发智能体，负责 React 应用构建
+- [[Rapid Prototyper]]：快速原型智能体，加速初始版本落地
+- [[Growth Hacker]]：增长黑客智能体，负责 Launch 策略规划
+- [[Reality Checker]]：现实检查智能体，在关键节点进行生产就绪评估和 GO/NO-GO 决策
+
+## Connections
+- [[Sprint Prioritizer]] → 输出冲刺计划 → [[Backend Architect]]
+- [[UX Researcher]] → 输出研究简报 → [[Backend Architect]]
+- [[Backend Architect]] → 输出 API Spec → [[Frontend Developer]]
+- [[Frontend Developer]] + [[Rapid Prototyper]] → 并行构建 → Week 2 输出
+- [[Reality Checker]] ← 中点评估 ← [[Frontend Developer]]（Week 2）
+- [[Frontend Developer]] → 继续开发 → [[Growth Hacker]]（Week 3 并行）
+- [[Reality Checker]] ← 最终评估 ← [[Growth Hacker]]（Week 4）
+- [[Reality Checker]] ← 生产就绪检查 ← RetroBoard MVP
+
+## Contradictions
+- 与单兵作战开发模式冲突：
+  - 冲突点：传统 MVP 开发通常由单一开发者完成，多智能体协作增加了交接成本
+  - 当前观点：多智能体协作通过专业化分工提升效率，通过质量门控确保交付质量
+  - 对方观点：小型 MVP 项目使用多智能体可能过度设计，顺序交接增加上下文丢失风险
--- a/wiki/sources/workflow-with-memory.md
+++ b/wiki/sources/workflow-with-memory.md
@@ -0,0 +1,57 @@
+---
+title: "Multi-Agent Workflow: Startup MVP with Persistent Memory"
+type: source
+tags: [multi-agent, workflow, memory, startup, mvp]
+date: 2026-04-21
+---
+
+## Source File
+- [[raw/Agent/agency-agents/examples/workflow-with-memory.md]]
+
+## Summary
+- 核心主题：基于 MCP Memory Server 的多智能体协作工作流，解决手动交接上下文丢失问题
+- 问题域：传统工作流中 agent 间交接依赖人工复制粘贴，会话超时、多 agent 共享上下文、QA 失败回滚、跨日项目等场景下的上下文断裂
+- 方法/机制：通过 MCP Memory Server 实现 `remember`、`recall`、`rollback`、`search` 操作，使 agents 自动存取项目状态，消除手动交接
+- 结论/价值：上下文持久化 → 无复制粘贴交接 → 跨会话连续性 → QA 回滚自动化 → 多日项目可管理
+
+## Key Claims
+- MCP Memory Server 可实现 agent 间自动上下文存取，消除手动复制粘贴交接
+- 记忆标签系统（tag everything with project name）是实现 recall 的关键
+- Reality Checker 因所有 agent 存储工作于 memory，获得完整项目可见性
+- Rollback 机制替代手动 undo，使 QA 失败恢复更紧密
+
+## Key Quotes
+> "With an MCP memory server installed, agents store their deliverables in memory and retrieve what they need automatically." — 核心价值主张
+> "Agents recall what they need automatically" vs "Copy-paste full output between agents" — 前后对比
+> "Tag everything with the project name: This is what makes recall work." — 实施关键模式
+
+## Key Concepts
+- [[MCP Memory Server]]：模型上下文协议记忆服务器，支持 remember/recall/rollback/search 操作
+- [[Multi-Agent Workflow]]：多智能体协作流水线，多角色 agents 顺序/并行执行
+- [[Memory Tagging]]：记忆标签系统，通过项目名和接收 agent 名标记记忆实现精准召回
+- [[Rollback Mechanism]]：回滚机制，agent 可回退到最后检查点并修复问题
+- [[Reality Checker]]：真实性检查智能体，在每个里程碑前评估可发货性和风险
+
+## Key Entities
+- [[Sprint Prioritizer]]：冲刺规划智能体，将项目拆分为 4 个每周冲刺
+- [[UX Researcher]]：用户体验研究智能体，进行竞品分析输出研究简报
+- [[Backend Architect]]：后端架构智能体，设计数据库 schema 和 REST API
+- [[Frontend Developer]]：前端开发智能体，基于 API spec 构建 React 应用
+- [[Rapid Prototyper]]：快速原型智能体，加速初始版本上线
+- [[Growth Hacker]]：增长黑客智能体，规划产品发布策略
+- [[RetroBoard]]：团队回顾工具名称，作为示例项目贯穿全文
+
+## Connections
+- [[Sprint Prioritizer]] → produces → [[Sprint Plan]]
+- [[UX Researcher]] → produces → [[Research Brief]]
+- [[Backend Architect]] → recalls → [[Sprint Plan]] + [[Research Brief]]
+- [[Backend Architect]] → produces → [[API Spec]] + [[Database Schema]]
+- [[Frontend Developer]] → recalls → [[API Spec]] + [[Database Schema]]
+- [[Growth Hacker]] → recalls → [[Reality Checker Verdict]] + [[Launch Plan]]
+- [[Reality Checker]] → recalls → ALL deliverables for full project visibility
+
+## Contradictions
+- 与 [[workflow-startup-mvp.md]] 冲突：
+  - 冲突点：相同项目（RetroBoard）但有无 MCP Memory Server 的两种版本
+  - 当前观点：带 Memory 的版本通过自动 recall 消除手动交接
+  - 对方观点：无 Memory 版本依赖人工复制粘贴上下文