๐Ÿง™ Skill Crafting Cheat Sheet

Core Mission: Define, structure, and validate custom agent skills (SKILL.md) to guide autonomous agents with 100% execution precision.

โ”Œโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”
โ”‚                        ๐Ÿง™ SKILL-CRAFTING CONTROL DASHBOARD             โ”‚
โ”œโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ฌโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ฌโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ค
โ”‚ ๐Ÿ“ FRONTMATTER    โ”‚ ๐Ÿงผ EXECUTION PHASES  โ”‚ โšก RULES & ANTI-PATTERNS    โ”‚
โ”œโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ผโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ผโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ค
โ”‚ 1. Name & Desc    โ”‚ 1. Cleansing         โ”‚ 1. Invariants               โ”‚
โ”‚ 2. Conditions     โ”‚ 2. Reflection        โ”‚ 2. Avoid walls of prose     โ”‚
โ”‚ 3. Specific paths โ”‚ 3. Verification      โ”‚ 3. Visual checks            โ”‚
โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ดโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜

๐Ÿ“ 1. METADATA & FRONTMATTER (Trigger Definition)

  • Trigger YAML Frontmatter: Every skill must begin with frontmatter defining metadata:
    • name: Short kebab-case name of the skill.
    • description: 1-2 sentence functional description.
    • condition: Semantic query triggers or regex-like patterns matching the user intent.
  • Condition Invariant: Ensure the triggers are highly specific to avoid false triggers or cross-trigger contamination.

๐Ÿงผ 2. EXECUTION PHASES (Standard Checklist Gates)

Every skill must divide agent operations into clear chronological steps using checklist items - [ ]:

  1. Cleansing & Noise Stripping: Directives to ignore conversational pleasantries, logs, and irrelevant context.
  2. Reflection & Core Invariant Extraction: Listing strict codebase rules, tsconfig target limits, dependency boundaries, or Nginx mappings.
  3. Condensation & Formatting: Presenting high-density markdown tables, comparative lists, and ASCII maps.

โšก 3. SKILL.MD BOILERPLATE TEMPLATE

Use this standard blueprint layout when creating or updating custom agent skills:

---
name: [skill-name]
description: [Single-sentence functional overview]
condition: [Trigger condition matching intent]
---
 
# ๐Ÿš€ [EMOJI] [CAPITALIZED SKILL NAME]
 
> **Core Directive:** [Single sentence bold directive]
 
[ASCII Dashboard]
 
---
 
## ๐Ÿงผ 1. CLEANSING & NOISE STRIPPING
 
- [Clean parameters, conversational templates, logs]
 
## ๐Ÿง  2. CORE REFLECTION RULES
 
- [Strict DO/DON'T boundaries, typing limits, performance gates]
 
## โšก 3. PHASES OF WORK
 
- [ ] Phase 1: Cleansing & Setup
- [ ] Phase 2: Implementation & Tests
- [ ] Phase 3: Verification & Deployment
 
## ๐Ÿ“‹ 4. OUTPUT TEMPLATES
 
- [Comparative tables, key-value summaries, ASCII layouts]
 
## ๐Ÿšจ 5. ANTI-PATTERNS
 
- โŒ [PITFALLS / WHAT NOT TO DO]