Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
- Added metadata file `_meta.json` for improved package management. - Updated daily reflection process in `assets/cron/daily-reflection.md`.
What This Skill Does
Configures an OpenClaw agent as a Chief, an autonomous business operator with trust-based permissions, a structured knowledge bank, sub-agent delegation patterns, and scheduled reflection cycles. Trust levels per action category determine whether the agent proposes actions, acts and notifies, or acts fully independently. Knowledge accumulates across daily, weekly, and monthly reflection cycles and is shared selectively with delegated workers.
The trust progression system prevents agents from acting autonomously too early while giving them a defined path to earn broader permissions through demonstrated reliability.
When to use it
- Delegating customer email triage to a sub-agent while the Chief reviews drafts before delivery
- Running a weekly reflection to update trust levels after consistent task successes
- Maintaining per-client knowledge pages that workers reference when drafting proposals
- Setting up daily cron jobs to extract learnings and prune stale memory entries
- Escalating a task category from propose to notify after three consecutive successful completions
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: Handle my customer emails
- 1Spawn a worker to check the inbox and list unread emails with sender, subject, and preview
- 2Review the list and categorize emails by urgency and type
- 3Spawn workers to draft responses for routine emails using context from bank/ and tone from shared/style-guide.md
- 4Review each draft, fix tone issues, and flag sensitive emails requiring human approval
- 5Deliver summary with routine emails queued and flagged items surfaced for review
Routine emails drafted and queued, pricing-related email flagged for human approval with full context