router

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Kinhluan Skill Router

Kinhluan Skill 路由器

Dispatch to the correct kinhluan-skill using 3-layer cascade.

通过三层级联机制将请求分发至正确的kinhluan-skill。

Layer 1: Hook Hint (auto)

第一层：钩子提示（自动触发）

Check if

[skill-router]

hint appeared in context (UserPromptSubmit hook output). If yes: use the suggested skill directly.

检查上下文（UserPromptSubmit钩子输出）中是否出现

[skill-router]

提示。如果出现：直接使用建议的技能。

Layer 2: Memory Recall

第二层：记忆召回

Invoke: memory-recall
Query: "kinhluan skill routing [topic keywords from user message]"

If memory returns a routing pattern → use it.

Invoke: memory-recall
Query: "kinhluan skill routing [topic keywords from user message]"

如果记忆返回路由模式 → 使用该模式。

Layer 3: Intent Matching

第三层：意图匹配

Map user intent → skill:

将用户意图映射至对应技能：

Research Track

研究领域

Intent	Skill
Find papers, survey field	`kinhluan-skills:sota-survey`
Write research question	`kinhluan-skills:research-question`
Design experiment	`kinhluan-skills:research-design`
Write paper/chapter	`kinhluan-skills:paper-writing`
PhD thesis defense	`kinhluan-skills:defense-prep`
Write PhD proposal	`kinhluan-skills:phd-proposal`
Track experiments	`kinhluan-skills:experiment-tracking`
Self-review paper	`kinhluan-skills:internal-critique`
Track milestones/deadlines	`kinhluan-skills:milestone-tracker`
Choose publication venue	`kinhluan-skills:publication-strategy`

意图	技能
查找论文、调研领域	`kinhluan-skills:sota-survey`
撰写研究问题	`kinhluan-skills:research-question`
设计实验	`kinhluan-skills:research-design`
撰写论文/章节	`kinhluan-skills:paper-writing`
博士论文答辩准备	`kinhluan-skills:defense-prep`
撰写博士研究计划	`kinhluan-skills:phd-proposal`
跟踪实验进度	`kinhluan-skills:experiment-tracking`
论文自我评审	`kinhluan-skills:internal-critique`
跟踪里程碑/截止日期	`kinhluan-skills:milestone-tracker`
选择发表期刊/会议	`kinhluan-skills:publication-strategy`

Architecture Track

架构领域

Intent	Skill
C4 overview (which level?)	`kinhluan-skills:c4-model`
System context diagram	`kinhluan-skills:c4-level1-context`
Container/service diagram	`kinhluan-skills:c4-level2-container`
Component/internal diagram	`kinhluan-skills:c4-level3-component`
Class/ER diagram	`kinhluan-skills:c4-level4-code`
Architecture decision	`kinhluan-skills:architecture-decision-records`
Evolutionary/fitness functions	`kinhluan-skills:evolutionary-architecture`

意图	技能
C4概览（对应哪个层级？）	`kinhluan-skills:c4-model`
系统上下文图	`kinhluan-skills:c4-level1-context`
容器/服务图	`kinhluan-skills:c4-level2-container`
组件/内部结构图	`kinhluan-skills:c4-level3-component`
类图/ER图	`kinhluan-skills:c4-level4-code`
架构决策记录	`kinhluan-skills:architecture-decision-records`
演化架构/适配度函数	`kinhluan-skills:evolutionary-architecture`

DDD Track

DDD领域

Intent	Skill
Strategic DDD, bounded context	`kinhluan-skills:ddd-core`
Aggregate, entity, value object	`kinhluan-skills:ddd-tactical`
CQRS, event sourcing, saga	`kinhluan-skills:ddd-patterns`

意图	技能
战略级DDD、限界上下文	`kinhluan-skills:ddd-core`
聚合根、实体、值对象	`kinhluan-skills:ddd-tactical`
CQRS、事件溯源、Saga模式	`kinhluan-skills:ddd-patterns`

DevOps Track

DevOps领域

Intent	Skill
Docker, Dockerfile	`kinhluan-skills:docker-containerization`
Kubernetes, k8s	`kinhluan-skills:kubernetes-orchestration`
DORA metrics, delivery	`kinhluan-skills:dora-core`

意图	技能
Docker、Dockerfile	`kinhluan-skills:docker-containerization`
Kubernetes、k8s	`kinhluan-skills:kubernetes-orchestration`
DORA指标、交付效率	`kinhluan-skills:dora-core`

Strategy Track

战略规划领域

Intent	Skill
Validate problem/demand	`kinhluan-skills:problem-discovery`
Why build this?	`kinhluan-skills:why-strategic-rationale`
JTBD, product strategy	`kinhluan-skills:business-product-leadership`
Release tracking, diffusion	`kinhluan-skills:diffusion-release-tracking`
Competitive strategy	`kinhluan-skills:art-of-war-software-engineering`

意图	技能
验证问题/需求	`kinhluan-skills:problem-discovery`
为何开发该产品？	`kinhluan-skills:why-strategic-rationale`
JTBD、产品战略	`kinhluan-skills:business-product-leadership`
发布跟踪、扩散管理	`kinhluan-skills:diffusion-release-tracking`
竞争战略	`kinhluan-skills:art-of-war-software-engineering`

ML Track

机器学习领域

Intent	Skill
Federated learning + DQN	`kinhluan-skills:federated-learning-dqn`
Scheduling algorithms, MLFQ	`kinhluan-skills:scheduling-algorithms`

意图	技能
联邦学习+DQN	`kinhluan-skills:federated-learning-dqn`
调度算法、MLFQ	`kinhluan-skills:scheduling-algorithms`

Writing Track

写作领域

Intent	Skill
Vietnamese CS terms	`kinhluan-skills:vietnamese-cs-terminology`
Technical English, IEEE/ACM	`kinhluan-skills:technical-english-cs`
Vietnamese writing standards	`kinhluan-skills:vietnamese-writing-standard`

意图	技能
越南语计算机科学术语	`kinhluan-skills:vietnamese-cs-terminology`
技术英语、IEEE/ACM规范	`kinhluan-skills:technical-english-cs`
越南语写作规范	`kinhluan-skills:vietnamese-writing-standard`

Browser / Web Track

浏览器/网页领域

Intent	Skill
Chrome Extension, browser automation	`kinhluan-skills:browser-automation`
Web scraping, data extraction	`kinhluan-skills:browser-automation`
Playwright / Puppeteer / CDP	`kinhluan-skills:browser-automation`
Form filling, UI automation	`kinhluan-skills:browser-automation`

意图	技能
Chrome扩展、浏览器自动化	`kinhluan-skills:browser-automation`
网页爬取、数据提取	`kinhluan-skills:browser-automation`
Playwright/Puppeteer/CDP	`kinhluan-skills:browser-automation`
表单填充、UI自动化	`kinhluan-skills:browser-automation`

Dev Track

开发领域

Intent	Skill
Python project setup	`kinhluan-skills:python-development`
TypeScript/JavaScript	`kinhluan-skills:javascript-typescript`
Security audit	`kinhluan-skills:security-analysis`

意图	技能
Python项目搭建	`kinhluan-skills:python-development`
TypeScript/JavaScript	`kinhluan-skills:javascript-typescript`
安全审计	`kinhluan-skills:security-analysis`

Dispatch Protocol

分发协议

Identify matched skill from above table
Announce: "Using
```
<skill>
```
to <purpose>"
Invoke:
```
Skill(<matched-skill>)
```
Follow skill instructions

从上述表格中识别匹配的技能
告知用户：“正在使用
```
<skill>
```
来完成<目标>”
调用：
```
Skill(<matched-skill>)
```
遵循该技能的操作指引

Memory Integration

记忆集成

After completing a routed task:

If routing was non-obvious or user confirmed it worked → trigger
```
memory-store
```

Store pattern:

"[user intent keywords] → [skill-name] — worked for [context]"

Future recalls will shortcut Layer 1-3 matching

完成路由任务后：

如果路由过程不明确或用户确认任务执行有效 → 触发
```
memory-store
```

存储模式：

"[用户意图关键词] → [技能名称] — 在[上下文]中执行有效"

后续记忆召回将跳过第一至第三层的匹配流程

When Multiple Skills Match

多技能匹配场景

Pick most specific:

```
c4-level1-context
```
>
```
c4-model
```
(more specific wins)
```
ddd-tactical
```
>
```
ddd-core
```
(if aggregate/entity mentioned)
Ask user if genuinely ambiguous (>2 equally valid options)

选择最具体的技能：

```
c4-level1-context
```
>
```
c4-model
```
（更具体的技能优先）
```
ddd-tactical
```
>
```
ddd-core
```
（如果提及聚合根/实体）
如果确实存在歧义（超过2个同等有效的选项），询问用户确认