WenHub

v1.5.0

文枢三院制 AI Agent 协作治理体系。提供决策者→管理者→经略院(任务拆解)→工造院(执行)→明鉴院(质检) 的完整协作流程,适用于需要多 Agent 协作治理的场景, 包括任务分配与分级、质量管控与评分、标准化汇报、知识沉淀与复用、安全管控与违规处理等。 配套参考资料(references/ 目录)涵盖三...

0· 104·0 current·0 all-time
by巩旭@18505298102

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for 18505298102/wenhub.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "WenHub" (18505298102/wenhub) from ClawHub.
Skill page: https://clawhub.ai/18505298102/wenhub
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Bare skill slug

openclaw skills install wenhub

ClawHub CLI

Package manager switcher

npx clawhub@latest install wenhub
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
Purpose & Capability
Name, description, and included reference docs (three-courts, task-grading, quality, security, reporting, violation handling) all align with a multi-agent governance framework; there are no unrelated required binaries, env vars, or config paths that would be unexpected for this purpose.
Instruction Scope
SKILL.md directs the user/agent to initialize governance files in the workspace (.wenhub/.rules/, SOUL.md, AGENTS.md) and to follow the included reference docs. It also instructs obtaining a template package from an external URL (https://wenhub.huawen-inc.com). The instructions do not ask for secrets, system config reads, or broad access beyond the workspace, but creating files and fetching templates are actions to be aware of.
Install Mechanism
No install spec, no code files, and no archive downloads are performed by the skill itself. The only external reference is a vendor 'download templates' URL mentioned in the docs (not an automated installer). Because nothing is automatically downloaded or extracted by the skill, install risk is low.
Credentials
The skill declares no required environment variables, credentials, or config paths. The included security guidance explicitly warns against embedding secrets. The lack of requested credentials is proportionate to a governance/template skill.
Persistence & Privilege
Skill is instruction-only, always:false, and does not request permanent presence or elevated agent/system privileges. It does instruct adding files to the workspace which is normal for a template/governance framework.
Assessment
This skill appears coherent for setting up a multi-agent governance workflow. Before using it: (1) Inspect any template package you download from the referenced URL before placing files into your workspace; don't run unknown installers. (2) Keep secrets out of the created template files (.wenhub/.rules/, SOUL.md, AGENTS.md) and follow the included security-rules guidance (use env vars or encrypted stores for keys). (3) Prefer to add the templates to a private repo or branch first and review them with code review or pre-commit scanners. (4) If you need higher assurance, verify the vendor domain (wenhub.huawen-inc.com) independently and request a signed release or checksum for any template archive. If you want, I can flag all lines in the templates that might hold sensitive placeholders to avoid accidentally committing credentials.

Like a lobster shell, security has layers — review code before you run it.

latestvk97ca9m1kt21y35xpq2geztztd85edyy
104downloads
0stars
5versions
Updated 4d ago
v1.5.0
MIT-0

文枢 · WenHub 治理体系

文以启智,枢运于行

触发条件

当以下场景出现时,加载本技能:

  • 需要多 Agent 协作治理与任务分配
  • 需要建立质量管控与评分体系
  • 需要标准化汇报与沟通流程
  • 需要知识沉淀与复用机制
  • 需要安全管控与违规处理框架
  • 需要快速初始化团队治理配置(开箱即用)

快速入门(5 步启用)

  1. 初始化:参考本技能文档,在工作区创建 SOUL.md、AGENTS.md 及 .wenhub/.rules/ 目录下的规则文件,替换占位符为实际值。
  2. 启动三院:按照 references/three-courts.md 中的组织架构启动经略院、工造院、明鉴院。
  3. 执行任务:根据 references/task-grading.md 进行任务分级,按对应流程执行。
  4. 汇报结果:使用 references/reporting.md 中的标准模板进行汇报。

核心概念

三院制:一种多 Agent 协作治理模式。

角色职责一句话定义
决策者战略方向、经营决策、文化建设定战略、把方向
管理者战略执行、团队管理、质量把控管执行、抓落地
经略院任务拆解、标准化、制定验收标准(DoD)谋定而后动
工造院按标准和流程执行,产出交付物天工开物
明鉴院独立检验交付物,出具质检报告明察秋毫

协作流程:决策者 → 管理者 → 经略院(拆解+DoD) → 工造院(执行) → 明鉴院(质检) → 管理者(验证) → 决策者(汇报)

操作指南

分级任务

启动任务前,先读取 references/task-grading.md,按紧急/常规/复杂三级分类执行:

  • 🔴 紧急:直接派工造院,跳过标准化
  • 🟡 常规:经略院 → 工造院 → 明鉴院
  • 🟢 复杂:经略院多轮评审 → 工造院 → 明鉴院

Spawn 三院

  • 经略院:负责任务拆解,输出标准化 DoD
  • 工造院:按 DoD 执行,产出交付物
  • 明鉴院:独立质检,输出通过/退回结论

质量评分

references/quality-system.md 执行评分,采用 5 维度加权(准确性 40%/完整性 20%/规范性 20%/效率 10%/创新性 10%),S~D 五级评定。

汇报

使用 references/reporting.md 中的标准模板,确保准确性、完整性、格式规范。

References 文件索引

文件何时读取核心内容
references/three-courts.md建立组织架构时三院职责、汇报铁律、流程图、知识沉淀、模型选择
references/task-grading.md任务启动前任务分级、前置检查、执行管控、失败处理、知识复用
references/quality-system.md任务验收时评分维度、等级定义、评分操作指南
references/reporting.md汇报阶段汇报模板、自检清单、沟通原则、群聊礼仪
references/security-rules.md涉及敏感数据时安全等级、通用铁律、存储安全、Git 安全
references/violation-system.md处理违规时违规定义、类型、处理流程、扣分机制
references/karpathy-rules.md编码任务时Karpathy 编码四原则、违反表现、团队执行要求

模板初始化参考(模板文件需从官网下载)

如需完整模板包(SOUL.md / AGENTS.md / .wenhub/.rules/ 等),请访问官网下载: https://wenhub.huawen-inc.com

模板文件清单:

文件目标位置用途
SOUL.md工作区根目录Agent 灵魂文件
AGENTS.md工作区根目录Agent 启动规则
IDENTITY.md工作区根目录身份指引
.wenhub/.rules/covenant.wenhub/.rules/公约框架
.wenhub/.rules/org-structure.wenhub/.rules/组织架构
.wenhub/.rules/security.wenhub/.rules/安全框架
.wenhub/.rules/etiquette.wenhub/.rules/礼仪框架
.wenhub/.rules/engineering.wenhub/.rules/工程框架
.wenhub/.rules/chanding.wenhub/.rules/禅定框架
.wenhub/.rules/violations.wenhub/.rules/违规框架
.wenhub/.journals/YYYY-MM-DD.md.wenhub/.journals/禅定日志

知识沉淀

任务完成 → 产出标准流程文档 → 存入知识库 → 同类任务优先检索复用。详见 references/three-courts.md 中的知识沉淀机制。

版本记录

版本日期说明
v1.0.02026-04-21初始版本(slug: wenhub-framework)
v1.1.02026-04-23更名为 wenhub;补齐 templates/ 目录;优化描述与结构

Comments

Loading comments...