Generate reusable prompts using the 7-section architecture for B2B SaaS marketing deliverables.
For full process flowchart and step-by-step → the premium reference.
Claude Code Triggers
Invoke this skill when user says:
"Create a prompt for [deliverable]"
"Build a prompt template"
"Generate a prompt for [task]"
"Prompt template for [use case]"
"Make me a prompt"
"Write a prompt for [deliverable]"
Do NOT invoke when:
User wants a workflow → Use workflow-design skill
User wants to execute a prompt → Just run it directly
User wants a specific skill output → Use the relevant skill
Input Requirements
Required
Input
Description
Source
Deliverable type
What output the prompt should produce
User specifies
Optional (improve quality)
Input
How it helps
Target audience
Tailors prompt language
Quality criteria
Defines success metrics
Existing prompts
Builds on proven patterns
Tool context
Adapts for specific AI (Claude, GPT, etc.)
If deliverable type is missing, ask the user before generating.
Core Framework: 7-Section Prompt Architecture
Every prompt follows this structure:
Section
Purpose
Content
ROLE
Sets expertise and context
"You are a [expert type] with [experience]..."
GOAL
System-level objective
"Your goal is to [outcome]..."
INPUTS
All {{variables}} user provides
List each variable with description
TASK
Step-by-step instructions
Numbered steps with clear actions
OUTPUT FORMAT
Structure, length, constraints
Format specification, character limits
CONTEXT
Reminder to pull from memory/knowledge
"Use any brand context, previous research..."
EXAMPLE
Placeholder for user's example
"Reference this example: {{example}}"
Prompt template
## ROLE
You are a [EXPERT TYPE] with deep expertise in [DOMAIN]. You have [SPECIFIC EXPERIENCE].
## GOAL
Your goal is to [PRIMARY OBJECTIVE] that [QUALITY CRITERIA].
## INPUTS
The user will provide:
- `{{variable_1}}`: [Description]
- `{{variable_2}}`: [Description]
## TASK
Follow these steps:
1. [First action with specific instruction]
2. [Second action with specific instruction]
3. [Third action with specific instruction]
## OUTPUT FORMAT
Deliver the output as:
- Format: [Markdown/JSON/etc.]
- Length: [Specification]
- Structure: [Description]
Include:
- [Required element 1]
- [Required element 2]
## CONTEXT
Pull from any available context:
- Brand guidelines and voice
- Previous research or deliverables
- Known audience information
## EXAMPLE
Reference this example for quality: {{example}}
Category matching
Match the user's request to the appropriate reference file:
Category
Use cases
Reference file
Research
Competitor, persona, win/loss
the premium reference
Social
LinkedIn, X, carousels
the premium reference
Long-form
Blog, articles
the premium reference
Landing pages
Homepage, persona, use case
the premium reference
Launch
Product announcements
the premium reference
Sales
Decks, battlecards
the premium reference
Positioning
Frameworks, canvases
the premium reference
Founder content
Thought leadership
the premium reference
AEO
Comparison, definition, how-to
the premium reference
Distribution
Channel distribution
the premium reference
Anti-Hallucination Guardrails
Use exact 7-section structure. Don't invent new sections.
Mark all variables. Every user input must be {{marked}}.
Include quality gates. Every prompt needs verification criteria.
Reference existing patterns. Check reference files before creating new.
Test variable completeness. Ensure all needed inputs are captured.
Integration with Other Skills
Skill
Relationship
Usage
workflow-design
Complementary
For multi-step workflows with chained prompts
All content skills
Uses output
Prompts feed into skill execution
vibe-coding
Related
For building tools that use prompts
1---2name: prompt-design3description: Prompt Design4---56# Prompt Design78Generate reusable prompts using the 7-section architecture for B2B SaaS marketing deliverables.910For full process flowchart and step-by-step → the premium reference.1112---1314## Claude Code Triggers1516**Invoke this skill when user says:**17- "Create a prompt for [deliverable]"18- "Build a prompt template"19- "Generate a prompt for [task]"20- "Prompt template for [use case]"21- "Make me a prompt"22- "Write a prompt for [deliverable]"2324**Do NOT invoke when:**25- User wants a workflow → Use `workflow-design` skill26- User wants to execute a prompt → Just run it directly27- User wants a specific skill output → Use the relevant skill2829---3031## Input Requirements3233### Required34| Input | Description | Source |35|-------|-------------|--------|36| **Deliverable type** | What output the prompt should produce | User specifies |3738### Optional (improve quality)39| Input | How it helps |40|-------|--------------|41| Target audience | Tailors prompt language |42| Quality criteria | Defines success metrics |43| Existing prompts | Builds on proven patterns |44| Tool context | Adapts for specific AI (Claude, GPT, etc.) |4546If deliverable type is missing, ask the user before generating.4748---4950## Core Framework: 7-Section Prompt Architecture5152Every prompt follows this structure:5354| Section | Purpose | Content |55|---------|---------|---------|56| **ROLE** | Sets expertise and context | "You are a [expert type] with [experience]..." |57| **GOAL** | System-level objective | "Your goal is to [outcome]..." |58| **INPUTS** | All `{{variables}}` user provides | List each variable with description |59| **TASK** | Step-by-step instructions | Numbered steps with clear actions |60| **OUTPUT FORMAT** | Structure, length, constraints | Format specification, character limits |61| **CONTEXT** | Reminder to pull from memory/knowledge | "Use any brand context, previous research..." |62| **EXAMPLE** | Placeholder for user's example | "Reference this example: {{example}}" |6364### Prompt template6566```markdown67## ROLE68You are a [EXPERT TYPE] with deep expertise in [DOMAIN]. You have [SPECIFIC EXPERIENCE].6970## GOAL71Your goal is to [PRIMARY OBJECTIVE] that [QUALITY CRITERIA].7273## INPUTS74The user will provide:75- `{{variable_1}}`: [Description]76- `{{variable_2}}`: [Description]7778## TASK79Follow these steps:801. [First action with specific instruction]812. [Second action with specific instruction]823. [Third action with specific instruction]8384## OUTPUT FORMAT85Deliver the output as:86- Format: [Markdown/JSON/etc.]87- Length: [Specification]88- Structure: [Description]8990Include:91- [Required element 1]92- [Required element 2]9394## CONTEXT95Pull from any available context:96- Brand guidelines and voice97- Previous research or deliverables98- Known audience information99100## EXAMPLE101Reference this example for quality: {{example}}102```103104### Category matching105106Match the user's request to the appropriate reference file:107108| Category | Use cases | Reference file |109|----------|-----------|----------------|110| Research | Competitor, persona, win/loss | the premium reference |111| Social | LinkedIn, X, carousels | the premium reference |112| Long-form | Blog, articles | the premium reference |113| Landing pages | Homepage, persona, use case | the premium reference |114| Launch | Product announcements | the premium reference |115| Sales | Decks, battlecards | the premium reference |116| Positioning | Frameworks, canvases | the premium reference |117| Founder content | Thought leadership | the premium reference |118| AEO | Comparison, definition, how-to | the premium reference |119| Distribution | Channel distribution | the premium reference |120121---122123## Anti-Hallucination Guardrails1241251. **Use exact 7-section structure.** Don't invent new sections.1262. **Mark all variables.** Every user input must be `{{marked}}`.1273. **Include quality gates.** Every prompt needs verification criteria.1284. **Reference existing patterns.** Check reference files before creating new.1295. **Test variable completeness.** Ensure all needed inputs are captured.130131---132133## Integration with Other Skills134135| Skill | Relationship | Usage |136|-------|--------------|-------|137| **workflow-design** | Complementary | For multi-step workflows with chained prompts |138| **All content skills** | Uses output | Prompts feed into skill execution |139| **vibe-coding** | Related | For building tools that use prompts |140141---142
Run npx skillmds@latest add matteotitta/prompt-design in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.
Prompt Design It is listed under AI & ML on SkillMD.
This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.
This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.
Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.
matteotitta (@matteotitta) published this skill. Their other Agent Skills are listed on their SkillMD profile.