Located in `.claude/agents/`:

**Hierarchy Level:** Reusable skill definitions

指导 Codex AI Agent 的工作方式。

Self-learning system that captures corrections during sessions and reminds users to run /reflect to update CLAUDE.md. Use when discussing learnings, corrections, or when the user mentions remembering something for future sessions.

'Turn a mature .agents corpus into packets, belief books, briefings, and gaps.'

'Plan an open source PR with scope, acceptance criteria, and risk.'

'Research an upstream repo before contributing: rules, patterns, and expectations.'

Agents are autonomous subprocesses that handle complex, multi-step tasks independently. Understanding agent structure, triggering conditions, and system prompt design enables creating powerful autonom

Plan, orchestrate, and validate deep research for nonfiction books. This skill

User flow design, micro-interactions, and interface behavior patterns with state management for all interaction conditions. Use when designing natural interactions, feedback systems, or handling error states.

Task and Event objects: polymorphic WhatId/WhoId, Activity object model, ActivityHistory vs OpenActivity, activity timeline customization, bulk task creation, Einstein Activity Capture boundaries. NOT for Calendar sharing (use calendar-sharing-setup). NOT for Email-to-Case (use case-management-setup).

Configure and manage Field Service Lightning work orders: work types, work order line items, service appointments, status flow, and auto-creation via maintenance plans. NOT for case management, resource scheduling optimization, or dispatcher console configuration.