inbox-action-board

What This Skill Does

The 'inbox-action-board' is a sophisticated productivity skill designed to transform chaotic inboxes and disorganized message streams into a structured, actionable dashboard. Instead of merely listing messages, this agent analyzes incoming communications and categorizes them into four distinct tactical views: 'Need to Reply,' 'Waiting for Others,' 'Needs Follow-up,' and 'Archiveable.' By applying GTD (Getting Things Done) principles, it helps users clear mental clutter and focus on high-impact tasks.

Installation

To install this skill, run the following command in your terminal: clawhub install openclaw/skills/skills/52yuanchangxing/inbox-action-board

Use Cases

• Email Triage: Paste a long list of email headers or threads to categorize them based on required user action.
• Project Management: Sync your chat messages from team channels to identify which tasks require you to chase stakeholders versus those requiring your direct input.
• Daily Planning: Use the 'What to do today' feature to extract high-priority action items from a backlog of mixed-media communications.
• Information Sanitization: Safely organize information without the risk of accidentally deleting original source data or performing unauthorized external actions.

Example Prompts

1. "I have 20 unread messages in my Slack. Please analyze the context and organize them into the four action categories for me."
2. "Review these email summaries from the last 24 hours. Create a prioritized action board, highlight anything I might have missed, and tell me what needs to be done today."
3. "Here is my mixed backlog of emails and notes. Create a draft actionable list, but identify any tasks that require me to wait on external responses from the marketing team."

Tips & Limitations

• Read-Only Philosophy: This skill is strictly designed for triage and organization. It will not perform destructive actions such as deleting original messages or sending emails on your behalf.
• Review First: Always treat the generated output as a draft. The agent is designed to provide an 'approvable plan' before you take any manual action.
• Data Privacy: If handling sensitive information, ensure you scrub PII before running the analysis if you are operating within an unsecure environment.
• Completeness: If the agent finds insufficient information, it will explicitly list 'Pending Confirmation' items rather than guessing your intent. Always check these sections to ensure no task falls through the cracks.