openclaw-executive-assistant-webinar

# OpenClaw Executive Assistant Webinar > Skill by [ara.so](https://ara.so) — Hermes Skills collection. ## Overview This project provides starter files and a structured workshop for building a local-first AI executive assistant using OpenClaw. It demonstrates three core workflows: 1. **Data intake review** – Turn unknown files into trustworthy intake reports 2. **Operational memory** – Transform work residue into daily logs and weekly summaries 3. **Offline communications triage** – Convert exported emails into actionable lists All workflows are local-only, produce reviewable markdown artifacts, and use copy/paste prompts with no live integrations. ## Repository Structure ``` . ├── webinar-runbook.html # Main workshop walkthrough └── code-along/ ├── INDEX.md ├── 01-data-intake-review/ │ ├── incoming/ # Files to inspect │ ├── prompts/intake-review.md # Report generation instructions │ ├── outputs/ # Generated reports │ └── expected/report-outline.md ├── 02-operational-memory/ │ ├── inbox/ # Work notes and residue │ ├── prompts/daily-log.md # Daily log prompt │ ├── prompts/weekly-hype.md # Weekly summary prompt │ ├── outputs/ # Generated logs │ └── schedule/ # Cron examples ├── 03-offline-communications-triage/ │ ├── eml/ # Exported email files │ ├── prompts/email-triage.md # Triage instructions │ ├── outputs/ # Triage reports │ └── expected/report-outline.md └── mission-control/ # Optional dashboard ``` ## Getting Started ### Installation ```bash git clone https://github.com/dandenney/webinars-build-your-own-executive-assistant-with-openclaw.git cd webinars-build-your-own-executive-assistant-with-openclaw ``` ### Workshop Flow 1. Open `webinar-runbook.html` in a browser 2. Keep the `code-along/` folder visible in your editor 3. Work through exercises sequentially 4. Copy prompts from `prompts/` directories 5. Review generated artifacts in `outputs/` directories ## Exercise 1: Data Intake Review **Goal:** Transform unknown incoming files into a structured intake report. ### File Structure ``` 01-data-intake-review/ ├── incoming/ # Place files to review here ├── prompts/ │ └── intake-review.md ├── outputs/ │ └── intake-review.md # Generated report └── expected/ └── report-outline.md ``` ### Usage Pattern 1. Place files to review in `incoming/` 2. Read the prompt from `prompts/intake-review.md` 3. Provide the prompt and file context to your AI assistant 4. Generate `outputs/intake-review.md` ### Expected Output Format The intake review should produce a markdown report containing: - **File inventory** – List of all files with types and sizes - **Content summary** – Brief description of each file's purpose - **Risk assessment** – Security/privacy concerns - **Recommended actions** – Next steps for each file - **Priority ranking** – Ordered by urgency/importance ## Exercise 2: Operational Memory **Goal:** Create daily logs and weekly summaries from work residue. ### File Structure ``` 02-operational-memory/ ├── inbox/ # Work notes, snippets, residue ├── prompts/ │ ├── daily-log.md │ └── weekly-hype.md ├── outputs/ │ ├── daily-log.md │ └── weekly-hype.md └── schedule/ ├── cron-examples.md └── heartbeat-note.md ``` ### Daily Log Pattern 1. Collect work residue in `inbox/` 2. Use `prompts/daily-log.md` to generate a daily log 3. Output to `outputs/daily-log.md` **Daily log structure:** - Date header - Completed tasks - In-progress work - Blockers/questions - Tomorrow's focus ### Weekly Summary Pattern 1. Accumulate daily logs over the week 2. Use `prompts/weekly-hype.md` to generate a weekly summary 3. Output to `outputs/weekly-hype.md` **Weekly summary structure:** - Week range header - Key accomplishments - Metrics/progress - Challenges addressed - Next week priorities ### Automation with Cron Reference `schedule/cron-examples.md` for automation patterns: ```bash # Daily log generation (5 PM weekdays) 0 17 * * 1-5 /path/to/generate-daily-log.sh # Weekly summary (Friday 5 PM) 0 17 * * 5 /path/to/generate-weekly-summary.sh ``` ## Exercise 3: Offline Communications Triage **Goal:** Convert exported email files into an actionable triage report. ### File Structure ``` 03-offline-communications-triage/ ├── eml/ # Exported .eml files ├── prompts/ │ └── email-triage.md ├── outputs/ │ └── email-triage.md # Generated triage └── expected/ └── report-outline.md ``` ### Usage Pattern 1. Export emails as `.eml` files into `eml/` 2. Use `prompts/email-triage.md` with your AI assistant 3. Generate `outputs/email-triage.md` ### Expected Triage Format The email triage report should contain: - **Urgent actions** – Emails requiring immediate response - **This week** – Items to address within 5 business days - **Backlog** – Lower-priority or FYI items - **Archive candidates** – No action needed - **Summary counts** – Total emails by category Each email entry should include: - Sender - Subject - Date received - Recommended action - Priority level ## Key Principles ### Local-First Architecture All data stays on your machine: - No cloud uploads - No API calls to external services - Reviewable markdown outputs - Version-controllable artifacts ### Copy/Paste Workflow 1. Navigate to exercise directory 2. Copy prompt from `prompts/*.md` 3. Paste into AI assistant (Claude, ChatGPT, etc.) 4. Provide file context as needed 5. Review and save output to `outputs/` ### Markdown Artifacts All outputs are markdown for: - Easy version control with Git - Plain-text searchability - Cross-platform compatibility - Human readability ## Common Patterns ### Adding Custom Prompts Create new prompt files following the structure: ```markdown # [Task Name] ## Context [What you're working with] ## Goal [What you want to produce] ## Instructions [Step-by-step guidance] ## Output Format [Expected structure] ``` ### Chaining Workflows Combine exercises for compound workflows: ```bash # 1. Review incoming files # outputs/intake-review.md # 2. Log the review work # outputs/daily-log.md (includes intake work) # 3. Triage any emails found # outputs/email-triage.md ``` ### Customizing Output Formats Edit prompt files to adjust output structure: - Change heading levels - Add custom sections - Modify priority categories - Include additional metadata ## Troubleshooting ### Missing Expected Output **Issue:** AI generates different format than expected **Solution:** Reference `expected/*.md` files to see the target structure, then refine your prompt with specific format requirements. ### File Context Too Large **Issue:** Too many files to process at once **Solution:** - Break into batches - Process high-priority files first - Create summary reports for large sets ### Inconsistent Daily Logs **Issue:** Daily logs vary in format day-to-day **Solution:** - Keep prompt files consistent - Use the same AI model - Reference previous logs as examples - Create a template in the prompt ### Cron Jobs Not Running **Issue:** Automated generation fails **Solution:** - Check cron syntax with `crontab -l` - Verify script paths are absolute - Ensure scripts have execute permissions: `chmod +x script.sh` - Check logs in `/var/log/cron` or system journal ## Best Practices 1. **Review all AI output** – Never blindly accept generated reports 2. **Version control artifacts** – Commit outputs to track changes over time 3. **Iterate on prompts** – Refine instructions based on output quality 4. **Keep raw inputs** – Preserve original files alongside processed outputs 5. **Regular cleanup** – Archive old outputs to maintain focus ## Integration Ideas While this workshop is local-only, you can extend it with: - File watching scripts to auto-trigger processing - Static site generation from markdown outputs - Notification systems when new reports are ready - Dashboard aggregation in `mission-control/` - Integration with note-taking tools (Obsidian, Logseq) ## Related Resources - DataCamp webinar: https://www.datacamp.com/webinars/build-your-own-executive-assistant-with-openclaw - OpenClaw documentation (check project homepage) - Markdown syntax reference: https://www.markdownguide.org/