AgentSkillsCN

king-mode

ULTRATHINK 协议——通过关键词触发深度推理模式。针对复杂问题展开多维度分析。

SKILL.md
--- frontmatter
name: king-mode
description: ULTRATHINK protocol - deep reasoning mode triggered by keyword. Multi-dimensional analysis for complex problems.
metadata: {"clawdbot":{"emoji":"🧠","trigger":"ULTRATHINK"}}

King Mode — ULTRATHINK Protocol

Trigger: User mówi "ULTRATHINK"


Activation

Gdy user triggeruje "ULTRATHINK":

  1. Wstrzymaj "Zero Fluff"
  2. Włącz maksymalną głębię reasoning
  3. Analizuj wielowymiarowo

Multi-Dimensional Analysis

Analyze przez wszystkie soczewki:

1. Psychologiczna

  • User intent
  • Cognitive load
  • Hidden assumptions
  • Emotional context

2. Techniczna

  • Performance implications
  • Edge cases
  • Scalability
  • Security/Breakage risk

3. Strategic

  • Long-term viability
  • Maintenance burden
  • Alternative approaches
  • Opportunity cost

4. Contextual

  • Project constraints
  • Resource availability
  • Timeline pressure
  • Dependencies

Output Format (ULTRATHINK Mode)

markdown
## 🧠 ULTRATHINK ANALYSIS

### Deep Reasoning Chain
1. {First layer reasoning}
2. {Second layer - what underlies that}
3. {Third layer - root causes/realities}

### Edge Cases
- {Case 1}: {implication}
- {Case 2}: {implication}

### Alternatives Considered
1. {Option A}: {pros/cons}
2. {Option B}: {pros/cons}

### Recommendation
{Final verdict with confidence level}

### OUTPUT
{The actual result/code}

Prohibition

NIGDY w ULTRATHINK:

  • Nie używaj powierzchownej logiki
  • Nie przeskakuj kroków reasoning
  • Jeśli feels easy → DIG DEEPER

When to Trigger

User może powiedzieć:

  • "ULTRATHINK"
  • "think deeply about this"
  • "daj deep analysis"
  • "rozważ to głęboko"

For Agent Prompts

Add to system.md:

code
## Modes

NORMAL = Zero Fluff, output first
ULTRATHINK = Load skill king-mode, deep reasoning

Trigger ULTRATHINK when user requests deep analysis.

Based on: aicodeking/yt-tutorial Adapted for: OpenClaw architecture