Claude Desktop - SKILL prompt

The Comprehensive Claude Skill Creation Prompt
Goal: Analyze the attached full conversation log (SKILL.md) to extract the essential knowledge, project goals, and constraints, and then synthesize this information into a structured, high-quality Claude Skill file.
Step 1: Core Knowledge Extraction and Synthesis
Action: Read the entire attached file. Do not summarize the conversation; instead, extract and synthesize the following structured data points:
Project Name: What is the primary, overarching project or topic of the conversation? (e.g., "DUI SEO Keyword Research Guide").
Skill Purpose: What is the specific, reusable function of this knowledge? (e.g., "Act as a strategic consultant for the DUI SEO project").
Key Deliverables/Outputs: List the concrete, final products that resulted from the conversation (e.g., "42 City Page Templates," "Pillar-Silo Architecture").
Constraints/Assumptions: Identify any specific tools, platforms, or limitations mentioned (e.g., "Implementation on Zizta.com," "Use of InfraNodus MCP").
Target User: Who is the assumed user of this Skill? (e.g., "SEO Manager," "Content Strategist").
Step 2: Skill Metadata Generation
Action: Based on the synthesis from Step 1, generate the required YAML front matter for the Skill file.
name: Use a concise, hyphenated name based on the Skill Purpose (e.g., dui-seo-strategy-consultant).
description: Write a single, clear sentence summarizing the Skill's knowledge and purpose.
license: Set to MIT (or another appropriate license if discussed in the chat).
allowed-tools: Set to [] unless the chat explicitly details a specific tool to be used.
metadata: Set to {}.
Step 3: Skill Instruction Set Development
Action: Create a detailed Instructions section for the Skill. This is the most critical part, as it tells Claude how to use the knowledge, not just what the knowledge is.
Role Definition: Start with a sentence defining the persona Claude should adopt when using this Skill (e.g., "You are now the DUI SEO Strategy Consultant.").
Knowledge Application: Describe the three primary ways Claude should apply the knowledge:
Strategic Alignment: Always cross-reference new requests with the original Project Goals and Constraints.
Detail Recall: When asked for specific items (e.g., "list of cities," "service names"), retrieve the exact content from the original conversation.
Future Planning: Use the Key Deliverables/Outputs as a checklist for future tasks and next steps.
Example Scenario: Provide one clear example of a user query and the expected, structured response from Claude using the Skill's knowledge.
Step 4: Final Output Generation
Action: Combine the generated metadata and the instruction set into a single, complete Markdown file.
Output Format:
Markdown
---
name: [Generated Skill Name]
description: [Generated Skill Description]
license: MIT
allowed-tools: []
metadata: {}
---

# Instructions for the [Generated Skill Name] Skill

## Role Definition
[Generated Role Definition]

## Knowledge Application Directives
1.  **Strategic Alignment:** [Directive 1]
2.  **Detail Recall:** [Directive 2]
3.  **Future Planning:** [Directive 3]

## Example Scenario
**User Query:** [Example Query]
**Expected Skill Response:** [Expected Response]

## Extracted Core Knowledge (For Reference)
| Field | Extracted Value |
| :--- | :--- |
| Project Name | [Value] |
| Key Deliverables | [Value] |
| Constraints/Platforms | [Value] |
| Target User | [Value] |
By following this four-step process, you are prompting Claude to perform the necessary analysis and structuring required to create a formal, reusable Skill, rather than just a memory summary. This ensures the output is a well-defined asset you can use for the Skill feature in Claude Desktop.