Back to Details

openclaw-executive-assistant-local

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

OpenClaw Executive Assistant (Local)

OpenClaw 本地执行助理

Skill by ara.so — Hermes Skills collection.
由ara.so提供的技能——Hermes技能合集。

Overview

概述

This project provides a local-first OpenClaw workflow system for building AI-powered executive assistant capabilities. It focuses on three core operations:
  1. Data intake review - Transform unknown files into structured intake reports
  2. Operational memory - Convert work residue into daily logs and weekly summaries
  3. Offline communications triage - Process exported emails into action lists
All operations use local files only, produce reviewable markdown artifacts, and require no live integrations.
本项目提供一个本地优先的OpenClaw工作流系统，用于构建AI驱动的执行助理能力。它专注于三个核心操作：
  1. 数据收集审核 - 将未知文件转换为结构化的收集报告
  2. 操作记忆 - 将工作残留内容转换为每日日志和每周总结
  3. 离线通讯分类处理 - 将导出的邮件处理为行动清单
所有操作仅使用本地文件，生成可审核的Markdown产物，无需实时集成。

Repository Structure

仓库结构

code-along/
├── INDEX.md
├── 01-data-intake-review/
│   ├── incoming/          # Files to inspect
│   ├── prompts/           # Prompt templates
│   ├── outputs/           # Generated reports
│   └── expected/          # Reference outputs
├── 02-operational-memory/
│   ├── inbox/             # Notes and work residue
│   ├── prompts/           # Daily/weekly prompts
│   ├── outputs/           # Generated logs
│   └── schedule/          # Cron examples
├── 03-offline-communications-triage/
│   ├── eml/               # Exported email files
│   ├── prompts/           # Triage prompts
│   ├── outputs/           # Triage reports
│   └── expected/          # Reference outputs
└── mission-control/       # Dashboard (optional)
code-along/
├── INDEX.md
├── 01-data-intake-review/
│   ├── incoming/          # 待检查文件
│   ├── prompts/           # 提示词模板
│   ├── outputs/           # 生成的报告
│   └── expected/          # 参考输出
├── 02-operational-memory/
│   ├── inbox/             # 笔记和工作残留内容
│   ├── prompts/           # 每日/每周提示词
│   ├── outputs/           # 生成的日志
│   └── schedule/          # Cron示例
├── 03-offline-communications-triage/
│   ├── eml/               # 导出的邮件文件
│   ├── prompts/           # 分类处理提示词
│   ├── outputs/           # 分类处理报告
│   └── expected/          # 参考输出
└── mission-control/       # 控制面板（可选）

Exercise 1: Data Intake Review

练习1：数据收集审核

Purpose

目标

Transform unknown files in an 
incoming/
folder into a trustworthy intake report.
将
incoming/
文件夹中的未知文件转换为可信的收集报告。

Workflow

工作流

  1. Place files to review:
bash
undefined
  1. 放置待审核文件：
bash
undefined

Add files to the incoming folder

将文件添加到incoming文件夹

cp ~/Downloads/unknown-file.pdf code-along/01-data-intake-review/incoming/

2. **Use the intake review prompt:**

The prompt template is in `prompts/intake-review.md`. Pass it to your AI assistant along with the contents of `incoming/`:

```markdown
cp ~/Downloads/unknown-file.pdf code-along/01-data-intake-review/incoming/

2. **使用收集审核提示词：**

提示词模板位于`prompts/intake-review.md`。将其与`incoming/`中的内容一起传递给你的AI助理：

```markdown

Intake Review Prompt

收集审核提示词

You are an executive assistant performing a data intake review.
Review all files in the incoming/ folder and produce a report that includes:
  1. File inventory (name, type, size, date)
  2. Content summary for each file
  3. Suggested categorization
  4. Action items or next steps
  5. Priority flags (urgent, routine, archive)
Output format: Markdown Output location: outputs/intake-review.md

3. **Expected output structure:**

```markdown
你是一名执行助理，正在进行数据收集审核。
审核incoming/文件夹中的所有文件并生成一份报告，内容包括：
  1. 文件清单（名称、类型、大小、日期）
  2. 每个文件的内容摘要
  3. 建议分类
  4. 行动项或下一步操作
  5. 优先级标记（紧急、常规、存档）
输出格式：Markdown 输出位置：outputs/intake-review.md

3. **预期输出结构：**

```markdown

Data Intake Review

数据收集审核

Generated: YYYY-MM-DD
生成时间：YYYY-MM-DD

Summary

摘要

  • Total files: X
  • Urgent items: Y
  • Requires action: Z
  • 文件总数：X
  • 紧急项：Y
  • 需要处理：Z

File Inventory

文件清单

[filename.ext]

[filename.ext]

  • Type: Document/Image/Data
  • Size: XXX KB
  • Date: YYYY-MM-DD
  • Summary: Brief content description
  • Category: Work/Personal/Archive
  • Action: Review/File/Respond
  • Priority: High/Medium/Low
[Repeat for each file]
  • 类型： 文档/图片/数据
  • 大小： XXX KB
  • 日期： YYYY-MM-DD
  • 摘要： 内容简要描述
  • 分类： 工作/个人/存档
  • 操作： 审核/归档/回复
  • 优先级： 高/中/低
[每个文件重复上述内容]

Recommended Actions

建议操作

  1. ...
  2. ...
undefined
  1. ...
  2. ...
undefined

Exercise 2: Operational Memory

练习2：操作记忆

Purpose

目标

Convert daily work residue into structured logs and weekly summaries.
将日常工作残留内容转换为结构化日志和每周总结。

Daily Log Workflow

每日日志工作流

  1. Add work residue to inbox:
bash
undefined
  1. 将工作残留内容添加到收件箱：
bash
undefined

Add notes, snippets, or quick captures

添加笔记、片段或快速记录

echo "Met with design team - new mockups ready" > code-along/02-operational-memory/inbox/note-$(date +%Y%m%d).txt

2. **Use the daily log prompt** (`prompts/daily-log.md`):

```markdown
echo "与设计团队会面 - 新原型已准备好" > code-along/02-operational-memory/inbox/note-$(date +%Y%m%d).txt

2. **使用每日日志提示词**（`prompts/daily-log.md`）：

```markdown

Daily Log Prompt

每日日志提示词

You are an executive assistant creating a daily work log.
Review all items in the inbox/ folder from today and produce:
  1. Date header
  2. Wins - Completed items
  3. Progress - Items in motion
  4. Blockers - Issues or delays
  5. Tomorrow - Planned next actions
  6. Notes - Observations or reminders
Output format: Markdown Output location: outputs/daily-log.md Filename pattern: daily-YYYY-MM-DD.md

3. **Expected daily log output:**

```markdown
你是一名执行助理，正在创建每日工作日志。
审核收件箱文件夹中今天的所有内容并生成：
  1. 日期标题
  2. 成果 - 已完成事项
  3. 进展 - 进行中事项
  4. 障碍 - 问题或延迟
  5. 明日计划 - 预计下一步操作
  6. 备注 - 观察或提醒
输出格式：Markdown 输出位置：outputs/daily-log.md 文件名格式：daily-YYYY-MM-DD.md

3. **预期每日日志输出：**

```markdown

Daily Log: 2026-05-11

每日日志：2026-05-11

Wins

成果

  • ✅ Completed data intake review system
  • ✅ Shipped v2.1 of client dashboard
  • ✅ 完成数据收集审核系统
  • ✅ 交付客户端仪表板v2.1

Progress

进展

  • 🔄 OpenClaw workshop prep (80% complete)
  • 🔄 Q2 planning document (draft stage)
  • 🔄 OpenClaw研讨会准备（完成80%）
  • 🔄 Q2规划文档（草稿阶段）

Blockers

障碍

  • ⚠️ Waiting on legal review for contract
  • ⚠️ Need API keys from DevOps
  • ⚠️ 等待合同的法律审核
  • ⚠️ 需要从DevOps获取API密钥

Tomorrow

明日计划

  • Finalize workshop slides
  • Review Q2 budget proposal
  • Team sync at 2pm
  • 最终确定研讨会幻灯片
  • 审核Q2预算提案
  • 下午2点团队同步会议

Notes

备注

  • Design team shared new mockups in Figma
  • Consider async standup format for remote team
undefined
  • 设计团队在Figma中分享了新原型
  • 考虑为远程团队采用异步站立会议格式
undefined

Weekly Summary Workflow

每周总结工作流

Use the weekly hype prompt (
prompts/weekly-hype.md
):
markdown
undefined
使用每周总结提示词（
prompts/weekly-hype.md
）：
markdown
undefined

Weekly Hype Prompt

每周总结提示词

You are an executive assistant creating a weekly summary.
Review all daily logs from this week (outputs/daily-*.md) and produce:
  1. Week of [date range]
  2. Highlights - Major wins and milestones
  3. Momentum - Projects advancing
  4. Attention needed - Recurring blockers
  5. Next week focus - Priorities for the week ahead
  6. Metrics (optional) - Quantifiable progress
Output format: Markdown Output location: outputs/weekly-hype.md Filename pattern: weekly-YYYY-Www.md
undefined
你是一名执行助理，正在创建每周总结。
审核本周所有每日日志（outputs/daily-*.md）并生成：
  1. [日期范围] 周
  2. 亮点 - 主要成果和里程碑
  3. 推进情况 - 进展中的项目
  4. 需要关注 - 反复出现的障碍
  5. 下周重点 - 本周优先事项
  6. 指标（可选）- 可量化的进展
输出格式：Markdown 输出位置：outputs/weekly-hype.md 文件名格式：weekly-YYYY-Www.md
undefined

Automation Example

自动化示例

Schedule daily log generation with cron (see 
schedule/cron-examples.md
):
bash
undefined
使用cron安排每日日志生成（参见
schedule/cron-examples.md
）：
bash
undefined

Run daily at 6pm

每天下午6点运行

0 18 * * * cd ~/openclaw-assistant && ./generate-daily-log.sh
0 18 * * * cd ~/openclaw-assistant && ./generate-daily-log.sh

generate-daily-log.sh example:

generate-daily-log.sh示例:

#!/bin/bash DATE=$(date +%Y-%m-%d) AI_PROMPT=$(cat code-along/02-operational-memory/prompts/daily-log.md)
#!/bin/bash DATE=$(date +%Y-%m-%d) AI_PROMPT=$(cat code-along/02-operational-memory/prompts/daily-log.md)

Pass inbox contents and prompt to your AI CLI tool

将收件箱内容和提示词传递给你的AI CLI工具

ai-cli "$AI_PROMPT" --context "code-along/02-operational-memory/inbox/*" \

ai-cli "$AI_PROMPT" --context "code-along/02-operational-memory/inbox/*" \

> "code-along/02-operational-memory/outputs/daily-$DATE.md"

> "code-along/02-operational-memory/outputs/daily-$DATE.md"

undefined
undefined

Exercise 3: Offline Communications Triage

练习3：离线通讯分类处理

Purpose

目标

Process exported email files into structured action lists.
将导出的邮件文件处理为结构化行动清单。

Workflow

工作流

  1. Export emails to .eml format:
bash
undefined
  1. 将邮件导出为.eml格式：
bash
undefined

Place exported emails in the eml/ folder

将导出的邮件放置在eml/文件夹中

cp ~/exported-emails/*.eml code-along/03-offline-communications-triage/eml/

2. **Use the email triage prompt** (`prompts/email-triage.md`):

```markdown
cp ~/exported-emails/*.eml code-along/03-offline-communications-triage/eml/

2. **使用邮件分类处理提示词**（`prompts/email-triage.md`）：

```markdown

Email Triage Prompt

邮件分类处理提示词

You are an executive assistant performing email triage.
Review all .eml files in the eml/ folder and produce:
  1. Urgent - Requires immediate response
  2. Action Required - Needs response (24-48h)
  3. FYI - Informational, no action needed
  4. Delegate - Should be handled by someone else
  5. Archive - Safe to file away
For each email include:
  • From/Subject
  • Brief summary
  • Suggested response or action
  • Priority level
Output format: Markdown Output location: outputs/email-triage.md

3. **Expected triage output:**

```markdown
你是一名执行助理，正在进行邮件分类处理。
审核eml/文件夹中的所有.eml文件并生成：
  1. 紧急 - 需要立即回复
  2. 需要处理 - 需要回复（24-48小时内）
  3. 仅供参考 - 信息性内容，无需操作
  4. 委派 - 应由其他人处理
  5. 存档 - 可安全归档
每封邮件需包含：
  • 发件人/主题
  • 简要摘要
  • 建议回复或操作
  • 优先级
输出格式：Markdown 输出位置：outputs/email-triage.md

3. **预期分类处理输出：**

```markdown

Email Triage Report

邮件分类处理报告

Generated: YYYY-MM-DD HH:MM
生成时间：YYYY-MM-DD HH:MM

Urgent (Response Today)

紧急（今日回复）

From: client@example.com | Re: Production Issue

发件人：client@example.com | 主题：生产环境问题

  • Summary: Database timeout errors affecting users
  • Action: Coordinate with DevOps for immediate fix
  • Priority: 🔴 Critical
  • 摘要： 数据库超时错误影响用户
  • 操作： 与DevOps协调立即修复
  • 优先级： 🔴 严重

Action Required (24-48h)

需要处理（24-48小时内）

From: legal@company.com | Re: Contract Review

发件人：legal@company.com | 主题：合同审核

  • Summary: Q2 vendor contract needs signature
  • Action: Review terms, sign if acceptable
  • Priority: 🟡 High
  • 摘要： Q2供应商合同需要签署
  • 操作： 审核条款，如可接受则签署
  • 优先级： 🟡 高

FYI (No Action)

仅供参考（无需操作）

From: team@company.com | Re: Weekly Newsletter

发件人：team@company.com | 主题：每周通讯

  • Summary: Company updates and team wins
  • Action: None - informational
  • Priority: 🟢 Low
  • 摘要： 公司更新和团队成果
  • 操作： 无 - 仅作信息参考
  • 优先级： 🟢 低

Delegate

委派

From: recruiter@agency.com | Re: Candidate Pipeline

发件人：recruiter@agency.com | 主题：候选人管道

  • Summary: Three candidates ready for interviews
  • Action: Forward to hiring manager Sarah
  • Priority: 🟡 Medium
  • 摘要： 三名候选人准备好面试
  • 操作： 转发给招聘经理Sarah
  • 优先级： 🟡 中

Archive

存档

[Emails that can be filed with no action]
undefined
[可归档无需操作的邮件]
undefined

Common Patterns

通用模式

Pattern 1: Copy-Paste Workflow

模式1：复制粘贴工作流

markdown
1. Open AI assistant (Claude, ChatGPT, etc.)
2. Copy prompt from prompts/*.md
3. Attach or paste relevant files from incoming/inbox/eml/
4. Run generation
5. Save output to outputs/*.md
6. Review and edit as needed
markdown
1. 打开AI助理（Claude、ChatGPT等）
2. 复制prompts/*.md中的提示词
3. 附加或粘贴incoming/inbox/eml/中的相关文件
4. 运行生成
5. 将输出保存到outputs/*.md
6. 根据需要审核和编辑

Pattern 2: Scripted Automation

模式2：脚本化自动化

bash
#!/bin/bash
bash
#!/bin/bash

automated-intake.sh

automated-intake.sh

PROMPT=$(cat code-along/01-data-intake-review/prompts/intake-review.md) FILES=$(ls code-along/01-data-intake-review/incoming/*)
PROMPT=$(cat code-along/01-data-intake-review/prompts/intake-review.md) FILES=$(ls code-along/01-data-intake-review/incoming/*)

Use your AI CLI tool of choice

使用你选择的AI CLI工具

ai-cli "$PROMPT" --files "$FILES" > outputs/intake-review-$(date +%Y%m%d).md

ai-cli "$PROMPT" --files "$FILES" > outputs/intake-review-$(date +%Y%m%d).md

undefined
undefined

Pattern 3: Scheduled Heartbeat

模式3：定时执行

bash
undefined
bash
undefined

Add to crontab

添加到crontab

Daily log at 6pm weekdays

工作日每天下午6点生成每日日志

0 18 * * 1-5 cd ~/openclaw-assistant && ./daily-log.sh
0 18 * * 1-5 cd ~/openclaw-assistant && ./daily-log.sh

Weekly summary Friday at 5pm

每周五下午5点生成每周总结

0 17 * * 5 cd ~/openclaw-assistant && ./weekly-summary.sh
undefined
0 17 * * 5 cd ~/openclaw-assistant && ./weekly-summary.sh
undefined

Configuration

配置

Environment Setup

环境设置

Create a 
.env
file for AI API configuration:
bash
undefined
创建
.env
文件用于AI API配置：
bash
undefined

.env

.env

ANTHROPIC_API_KEY=${ANTHROPIC_API_KEY} OPENAI_API_KEY=${OPENAI_API_KEY} AI_MODEL=claude-3-5-sonnet-20241022
undefined
ANTHROPIC_API_KEY=${ANTHROPIC_API_KEY} OPENAI_API_KEY=${OPENAI_API_KEY} AI_MODEL=claude-3-5-sonnet-20241022
undefined

Prompt Customization

提示词自定义

Edit prompt files to match your workflow:
bash
undefined
编辑提示词文件以匹配你的工作流：
bash
undefined

Customize intake review categories

自定义收集审核分类

nano code-along/01-data-intake-review/prompts/intake-review.md
nano code-along/01-data-intake-review/prompts/intake-review.md

Adjust daily log sections

调整每日日志章节

nano code-along/02-operational-memory/prompts/daily-log.md
nano code-along/02-operational-memory/prompts/daily-log.md

Modify email triage buckets

修改邮件分类处理分组

nano code-along/03-offline-communications-triage/prompts/email-triage.md
undefined
nano code-along/03-offline-communications-triage/prompts/email-triage.md
undefined

Troubleshooting

故障排除

Issue: AI output not matching expected format

问题：AI输出不符合预期格式

Solution: Add explicit format instructions to prompts:
markdown
CRITICAL: Output must be valid Markdown with exactly these sections:
- Summary
- File Inventory
- Recommended Actions

Use ## for section headers. Use - for bullet lists.
解决方案： 在提示词中添加明确的格式说明：
markdown
重要提示：输出必须为有效的Markdown，且包含以下章节：
- 摘要
- 文件清单
- 建议操作

使用##作为章节标题，使用-作为项目符号列表。

Issue: Large files causing context limits

问题：大文件导致上下文限制

Solution: Process in batches:
bash
undefined
解决方案： 分批处理：
bash
undefined

Split incoming files into chunks

将incoming文件拆分为多个批次

for file in incoming/*.pdf; do

Process individually

echo "Processing $file..." done
undefined
for file in incoming/*.pdf; do

单独处理

echo "正在处理$file..." done
undefined

Issue: Automation script not running

问题：自动化脚本未运行

Solution: Check cron logs and permissions:
bash
undefined
解决方案： 检查cron日志和权限：
bash
undefined

View cron logs

查看cron日志

grep CRON /var/log/syslog
grep CRON /var/log/syslog

Ensure scripts are executable

确保脚本可执行

chmod +x *.sh
chmod +x *.sh

Test script manually

手动测试脚本

./daily-log.sh
undefined
./daily-log.sh
undefined

Issue: Email .eml parsing errors

问题：邮件.eml解析错误

Solution: Ensure proper export format from email client. Most clients support "Save as .eml" or "Export to file" options. If parsing fails, extract plain text first:
bash
undefined
解决方案： 确保从邮件客户端正确导出格式。大多数客户端支持“另存为.eml”或“导出到文件”选项。如果解析失败，先提取纯文本：
bash
undefined

Extract text from .eml

从.eml中提取文本

grep -A 1000 "^$" email.eml | tail -n +2 > email.txt
undefined
grep -A 1000 "^$" email.eml | tail -n +2 > email.txt
undefined

Integration Tips

集成技巧

With Obsidian/Notion

与Obsidian/Notion集成

bash
undefined
bash
undefined

Symlink outputs to your notes folder

将输出目录符号链接到你的笔记文件夹

ln -s ~/openclaw-assistant/code-along/02-operational-memory/outputs ~/Obsidian/Daily-Logs
undefined
ln -s ~/openclaw-assistant/code-along/02-operational-memory/outputs ~/Obsidian/Daily-Logs
undefined

With Git for Versioning

使用Git进行版本控制

bash
undefined
bash
undefined

Track generated logs

跟踪生成的日志

cd code-along/02-operational-memory/outputs git init git add daily-.md weekly-.md git commit -m "Daily log archive"
undefined
cd code-along/02-operational-memory/outputs git init git add daily-.md weekly-.md git commit -m "每日日志存档"
undefined

With Markdown Viewers

与Markdown查看器集成

bash
undefined
bash
undefined

Serve outputs as local site

将输出作为本地站点提供

cd code-along python -m http.server 8000
cd code-along python -m http.server 8000

Open http://localhost:8000

打开http://localhost:8000

undefined
undefined

Best Practices

最佳实践

  1. Review before archiving - Always human-review AI outputs before filing
  2. Consistent naming - Use ISO date formats (YYYY-MM-DD) in filenames
  3. Regular cleanup - Archive old logs monthly to keep folders manageable
  4. Prompt iteration - Refine prompts based on output quality over time
  5. Local-first - Keep sensitive data local; only upload sanitized examples
  1. 存档前审核 - 在归档前始终人工审核AI输出
  2. 命名一致 - 文件名使用ISO日期格式（YYYY-MM-DD）
  3. 定期清理 - 每月归档旧日志以保持文件夹整洁
  4. 提示词迭代 - 根据输出质量随时间优化提示词
  5. 本地优先 - 敏感数据保持本地存储；仅上传经过清理的示例