菜单
Use this skill when you want to learn about generating high-quality insights that will be automatically captured. Insights using the "★ Insight" format are automatically extracted by hooks - no manual saving required.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Event-driven coordination for teamwork orchestrators. Replaces polling-based monitoring loop with TeammateIdle and TaskCompleted hooks.
Hybrid task decomposition strategy for teamwork projects. Covers plan-based and semantic decomposition, granularity rules, role assignment, dependency management, and complexity assessment.
Reset and clean up teamwork project state. Use when project is stuck, failed, or needs a fresh start. Uses native TeamDelete API for cleanup.
8-phase task execution lifecycle for teamwork workers using native Claude Code APIs. Covers task discovery, claiming, structured description parsing, TDD workflow, implementation, verification, selective commit, and completion reporting.
Architecture design review skill for reviewer agents. Component boundaries, separation of concerns, modularity, and design pattern validation. Use during PLANNING phase to validate design decisions and VERIFICATION to ensure adherence.
API design, database operations, error handling, and backend architecture patterns. Use when creating APIs, database schemas, or server-side logic.
| name | insight-awareness |
| description | Use this skill when you want to learn about generating high-quality insights that will be automatically captured. Insights using the "★ Insight" format are automatically extracted by hooks - no manual saving required. |
This skill guides you on creating valuable insights that will be automatically captured by hooks. You don't need to manually save insights - just generate them in the ★ Insight format and the Stop/SubagentStop hooks will extract them from the transcript.
1. You generate: ★ Insight ─────────────────────────────────────
[valuable knowledge]
─────────────────────────────────────────────────
2. Hook automatically:
- Parses transcript after your response
- Extracts content between markers
- Saves to ~/.claude/knowledge-extraction/{session-id}/insights.md
- Tracks state to avoid duplicates
Generate ★ Insight markers when you discover:
| Type | Description | Example |
|---|---|---|
code-pattern | Reusable patterns | "Prefer useReducer for complex form state" |
workflow | Efficient processes | "Run tests before commit hooks" |
debugging | Root cause findings | "Memory leak caused by unclosed listener" |
architecture | Design decisions | "Use event sourcing for audit trail" |
tool-usage | Effective techniques | "Combine Grep + Read for targeted searches" |
standard | Standards and conventions | "JSON files use 2-space indentation" |
convention | Naming and file patterns | "Scripts follow entity-action.js naming" |
Use the standard format for automatic extraction:
★ Insight ─────────────────────────────────────
[2-5 lines of valuable, reusable knowledge]
─────────────────────────────────────────────────
After insights accumulate:
/insights - View collected insights/insights extract - Convert to Skills/Commands/CLAUDE.md/Rules Files/insights clear - Clear session insights