LMI Management System

lmi-management-system is an OpenClaw / Codex skill that turns LMI-style role planning, time management, and review into a repeatable operating system.

It is designed for people who want an agent to help with:

• role goal clarification
• monthly planning
• weekly planning
• daily planning
• daily, weekly, monthly, and role review
• time design and personal productivity analysis

Why This Skill Exists

Many planning systems stop at templates. This skill is meant to go further:

• convert role goals into monthly, weekly, and daily action
• separate planning from review
• separate ideas from validated project facts
• force unfinished work to be re-decided instead of endlessly rolling forward
• help an agent act like a management collaborator, not just a form filler

What Is Included

The skill currently supports 10 working modes:

1. Role Goal Clarifier
2. Monthly Plan
3. Weekly Plan
4. Daily Review
5. Weekly Review
6. Monthly Review
7. Role Review
8. Daily Work Plan
9. Time Image
10. Stats Analysis

Core Ideas

• Goals come before tasks
• Time should follow goals
• High-return activities deserve protected time
• Role goals must flow into month, week, and day
• Reviews should create adjustments, not just observations

Planning And Review Loop

Use the skill as a closed loop:

Role Goal Clarifier
-> Monthly Plan
-> Weekly Plan
-> Daily Work Plan
-> Daily Review
-> Weekly Review
-> Monthly Review
-> Role Review

Correction loop:

Time Image
<-> Stats Analysis
-> Weekly Plan / Monthly Review

Daily Classification

For day-to-day execution, classify work like this:

• A: important items
• B: urgent items
• C: contact or follow-up items
• D: meetings, discussions, or coordination items

This is used to decide attention, not just to label tasks.

Recommended Storage Pattern

If you use an OpenClaw memory workspace, a practical storage layout is:

memory/
├── YYYY-MM-DD.md
├── 日报归档/YYYY-MM-DD_工作日志.md
├── 周复盘-YYYY-MM-DD-YY.md
├── 月复盘-YYYY-MM.md
├── 角色复盘-角色名-YYYY-MM.md
├── 项目事实/
└── 长期偏好/

Recommended usage:

• memory/YYYY-MM-DD.md: raw daily log and lightweight daily review
• 日报归档: cleaner daily log
• 周复盘 / 月复盘 / 角色复盘: formal reviews
• 项目事实: validated conclusions and stable project knowledge
• 长期偏好: stable working preferences and recurring rules

Management Rules

• Unfinished work must be re-decided each day: move, defer, delegate, or stop
• New work should be classified before entering the day plan
• Role goals should flow into month, week, and day instead of staying in a static role sheet
• Reflections belong in reviews
• Ideas should be captured first, then filtered weekly
• Only validated conclusions should be stored as project facts

Installation

OpenClaw

Install into the managed OpenClaw skills directory:

cp -R lmi-management-system ~/.openclaw/skills/

Then verify:

openclaw skills info lmi-management-system

Codex

Install into the Codex skills directory:

cp -R lmi-management-system ~/.codex/skills/

Quick Start Prompts

These prompts are good first tests:

• 用 LMI 帮我梳理这个岗位的角色目标澄清表
• 按 LMI 帮我做本月计划，先写使命/宗旨、个人焦点目标和公司焦点目标
• 用 LMI 把我的月目标拆成这周的工作计划
• 用 LMI 帮我排今天，按 A/B/C/D 分类，并给我日程安排
• 按 LMI 帮我做今天的日复盘，并给出明天第一步
• 按 LMI 周复盘帮我看：这周哪些是高回报活动，哪些时间浪费了
• 按 LMI 帮我做月复盘，检查个人目标、公司目标和机制问题
• 按 LMI 帮我做角色复盘，看看我的职责、关键业绩和时间投入是否匹配
• 按 LMI 时间图像帮我设计理想的一周时间分配
• 用 LMI 个人生产力摘要表分析我这周的时间投入

Example Output

Here is a short example of the kind of structured output this skill is meant to produce.

Prompt:

用 LMI 把我的月目标拆成这周的工作计划

Example output:

Week Theme
- 本周主题：把月目标收敛成可推进的 2 个结果
- 对应月目标：完成制造业场景验证，并建立客户跟进机制

This Week's Top Goals
- Goal 1：完成制造业场景 demo 的第一轮可演示版本
- Goal 2：完成 8 个重点客户的跟进触达

High-Return Activities
- Activity 1：保护 2 个 90 分钟深度工作块做 demo
- Activity 2：集中 1 个时间块完成客户跟进 SOP

Delegation And Follow-Up
- Must personally drive：demo 核心逻辑、客户价值主张
- Delegate：资料整理、会后纪要
- Follow up only：客户触达结果汇总

Next 1-3 Moves
- Move 1：先确定本周只保留 2 个 top goals
- Move 2：把周三前的深度工作时间块锁定
- Move 3：列出 8 个客户跟进名单和触达方式

Using With Azai

See references/azai-usage-guide.md for a practical usage guide and ready-to-send prompts for azai.

For OpenClaw / Feishu / agent-routing verification after changes, use:

• references/openclaw-feishu-lmi-regression-checklist.md

Recommended standing instruction:

以后请优先使用 lmi-management-system 协同我。默认先判断我是角色目标澄清、月计划、周计划、日计划、日复盘、周复盘、月复盘、角色复盘、时间图像还是统计分析场景。尽量先起草，再最少追问；优先压缩重点，默认检查高回报活动、授权事项和管理机制，并在结尾输出 Next 1-3 Moves。

Repository Layout

lmi-management-system/
├── SKILL.md
├── README.md
├── agents/
│   └── openai.yaml
├── assets/
│   └── templates/
├── references/
└── scripts/
    └── check-sync.sh

Important Files

• SKILL.md: main routing and workflow rules
• agents/openai.yaml: UI metadata
• assets/templates/: reusable planning and review templates
• references/: methodology, storage, review, and execution guidance
• scripts/check-sync.sh: checks whether source, OpenClaw, and Codex copies are still in sync

Main Files To Read

• SKILL.md
• CHANGELOG.md
• RELEASE_NOTES_v1.0.0.md
• references/storage-system.md
• references/openclaw-execution-playbook.md
• references/review-system.md
• references/azai-usage-guide.md

Current Status

This repository is the source-of-truth version used to sync:

• local development copy
• OpenClaw managed skill copy
• Codex skill copy

License

MIT. See LICENSE.

Repository

GitHub: woohuaca/lmi-management-system