/install feishu-work-review-timemachine
Feishu Work Review Timemachine
Overview
Create a compact, high-signal review of a user's recent work from Feishu materials. Focus on what the user actually drove or owned, not everything discussed in meetings.
Workflow
1. Define scope first
Clarify or infer:
- time range: last week / last month / specific dates
- output form: concise recap / weekly breakdown / monthly summary / boss-facing version
- destination: chat reply only, or create/update a Feishu doc
If the user already gave a clear scope, do not ask again.
2. Gather source materials from Feishu
Prefer this order:
- Search the user's recent docs/wiki nodes
- Prioritize docs created by or owned by the user
- Read key docs in full only when needed
- Prefer summaries / structured notes first, then raw transcript docs for validation
Useful source types:
- meeting notes and transcripts
- work-in-progress docs
- project plans
- technical方案 / design docs
- quality / metrics docs
- wiki pages tied to ongoing initiatives
Do not try to read everything. Read enough to identify repeated themes and evidence of ownership.
3. Separate real work from meeting noise
Do not treat every meeting mention as the user's work.
Use these signals to infer stronger ownership:
- the user is the doc owner / creator / primary editor
- the user is the main speaker framing goals, tradeoffs, or next steps
- action items are explicitly assigned to the user
- the user is pushing the plan, organizing follow-up, or coordinating POCs
- the same initiative appears across multiple docs with the user in a central role
Treat these as weaker signals:
- the user merely attended the meeting
- another person's project status appears in notes
- broad team discussion without assigned ownership
- AI-generated meeting summary claims not supported elsewhere
When uncertain, downshift wording:
- use “参与推动 / 协同推进 / 讨论并判断”
- avoid overstating as “主导完成”
4. Extract only the core themes
Compress the material into a few workstreams, usually 3-5:
- platform / product / project track
- AI / efficiency / technical exploration track
- metrics / governance / methodology track
- cross-team coordination track
For each workstream, capture only:
- what changed
- what the user specifically drove
- why it mattered
Delete or skip:
- optional commentary
- long caveats
- repeated explanations
- every subtask from meeting minutes
- alternate interpretations unless the user asked for analysis
5. Write in layers
Default output order:
- one-paragraph core conclusion
- 3-5 core workstreams
- weekly breakdown if requested
- one-sentence summary
If the user says “keep only the core”, reduce to:
- core conclusion
- core workstreams
- optional one-line weekly view
Output pattern
Compact monthly review
Use this structure:
# 时光机回顾|过去一月主要工作总结
## 核心结论
[1 short paragraph]
## 核心主线
### 1. [主线]
- [核心动作]
- [核心动作]
### 2. [主线]
- [核心动作]
## 按周回顾
### 第 1 周:[关键词]
[1-3 lines]
### 第 2 周:[关键词]
[1-3 lines]
## 一句话总结
[1 sentence]
Boss-facing version
Make it shorter and more outcome-oriented:
- progress
- scope / influence
- cross-team traction
- next-step signal
Avoid process-heavy wording.
Feishu doc creation/update
When the user asks to create or write a Feishu doc:
- create the doc once a solid draft exists
- if permissions block the action, tell the user to complete the authorization card and then resume
- once created, prefer overwriting the draft if the user asks to simplify aggressively
- when refining, keep the same doc unless the user asks for a new one
Style rules
- Write in concise Chinese unless the user asks otherwise
- Prefer short sections over long prose
- Remove “可选项” and hedging text unless uncertainty is material
- Use strong verbs for high-confidence ownership, softer verbs for medium-confidence involvement
- Default to fewer bullets, not more
- Keep the final document skimmable in under 1-2 minutes unless the user asked for depth
Anti-patterns
Do not:
- dump a list of all meetings
- copy AI meeting summaries verbatim
- equate attendance with ownership
- preserve every nuance when the user wants a crisp recap
- keep “next-step suggestions” if the user asked for a finished summary
Final check
Before delivering, verify:
- Is this mostly about the user's actual work, not everyone else's?
- Did I collapse repeated themes into a few clear workstreams?
- If asked to simplify, did I remove optional sections rather than just shorten sentences?
- Would the user immediately recognize this as accurate and useful?
- 确保已安装 OpenClaw(本地或 Docker 部署)
- 在对话框中输入安装命令:
/install feishu-work-review-timemachine - 安装完成后,直接呼叫该 Skill 的名称或使用
/feishu-work-review-timemachine触发 - 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
work-review-timemachine for feishu 是什么?
Summarize a user's recent Feishu work into a concise weekly review or "time machine" recap by reading Feishu docs, meeting notes, and related materials, then... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 274 次。
如何安装 work-review-timemachine for feishu?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install feishu-work-review-timemachine」即可一键安装,无需额外配置。
work-review-timemachine for feishu 是免费的吗?
是的,work-review-timemachine for feishu 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。
work-review-timemachine for feishu 支持哪些平台?
work-review-timemachine for feishu 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。
谁开发了 work-review-timemachine for feishu?
由 WangLei(@nihiue)开发并维护,当前版本 v1.0.0。