What Shall We Build Next?

Describe

Describe your task

Refine

Refine the plan

SubAgents

Review all agents

Deploy

Deploy your agent

Let me break this down for the "Niche Research & Pun Crafter" subagent:

A) SUBAGENT SUMMARY:
A specialized research and creative writing agent that analyzes a given niche, studies existing birthday card designs, and generates a curated list of niche-specific puns and slogans suitable for birthday cards.

B) FINAL TASK OUTPUT:
A text file ("pun_list.txt") containing 5-10 niche-specific birthday card puns/slogans, each formatted as a single line of text, with each pun being 2-15 words in length.

C) SUBAGENT INPUT:
- Primary niche/topic keyword (e.g., "gardening", "CrossFit", "programming")
- Optional secondary keywords or specific focus areas within the niche

E) SUBAGENT TASK SUMMARY:
The workflow breaks into three main components that chain together:

1. Market Research Pipeline:
Input: niche keyword >
#187 (Find Relevant Images) >
#176 (Analyze Images with GPT Vision) >
Output: market analysis text

2. Niche Research Pipeline:
Input: niche keyword >
#216 (Research Topic Deeply) >
#218 (Brainstorm Related Keywords) >
Output: comprehensive niche research text

3. Pun Generation Pipeline:
Input: combined research from pipelines 1+2 >
#223 (Powerful LLM) to synthesize research >
#186 (Create MicroCopy) x2 rounds >
#190 (Write/Rewrite Text) for final refinement >
Output: final pun_list.txt

F) SILOS:
SILO 1: MARKET RESEARCH
Purpose: Analyze existing birthday cards in the niche
- Skill #187: Search "<niche> birthday cards"
- Skill #176: Analyze each card image for:
  * Common visual themes
  * Existing pun patterns
  * Popular design elements
Output: Market analysis text document

SILO 2: NICHE RESEARCH
Purpose: Deep dive into niche terminology and culture
- Skill #216: Research niche deeply focusing on:
  * Common terminology
  * Inside jokes
  * Cultural references
- Skill #218: Generate related keywords to expand scope
Output: Niche research text document

SILO 3: PUN GENERATION
Purpose: Create and refine puns
- Skill #223: Synthesize research into creative brief
- Skill #186 (Round 1): Generate 10 initial puns
- Skill #186 (Round 2): Generate 10 more puns
- Skill #190: Refine and select best 5-10 puns
Output: Final pun_list.txt

This structure ensures a systematic approach from research to final output, with each silo building upon the previous one's work.

SubAgent #1 - Diagram

Expand Diagram

Let me break down the Image Prompt Generator subagent following your framework:

A) SUBAGENT SUMMARY: 
A specialized prompt-writing agent that converts birthday card puns into detailed image generation prompts, optimizing for style consistency and visual appeal across multiple cards.

B) FINAL TASK OUTPUT:
A text file containing 3-6 carefully crafted image generation prompts (200-400 characters each), each corresponding to a specific pun, with explicit style guidelines, composition details, and text placement instructions formatted for direct use with image generation AI.

C) SUBAGENT INPUT:
- List of 3-6 approved birthday card puns (from Subagent 1)
- Reference images/examples of successful birthday cards in the niche (from Subagent 1's research)

E) SUBAGENT TASK SUMMARY:
1. Analyze reference images:
#176 - Analyze An Image With GPT Vision & Return Text
INPUT: Top 3 reference images + "Analyze artistic style, composition, and text placement"
OUTPUT: Detailed style analysis text

2. Create style guide:
#190 - Write or rewrite text based on instructions
INPUT: Style analysis + "Create consistent style guide for birthday card series"
OUTPUT: Style guide text

3. Generate base prompts (3x parallel processes):
#190 - Write or rewrite text based on instructions
INPUT: Style guide + individual pun + "Create detailed image generation prompt"
OUTPUT: Initial prompt draft

4. Optimize prompts for image generation:
#223 - Powerful LLM Prompt-to-Text Response
INPUT: Initial prompts + "Optimize prompts for Dalle image generation"
OUTPUT: Final optimized prompts

F) SILOS:

SILO 1: STYLE ANALYSIS
- Input: Reference images
- Skill: #176 (Image analysis)
- Skill: #190 (Style guide creation)
- Output: Style guide text

SILO 2: PROMPT GENERATION (repeated 3-6 times, one per pun)
- Input: Individual pun + style guide
- Skill: #190 (Initial prompt writing)
- Skill: #223 (Prompt optimization)
- Output: Final prompt text

SILO 3: COMPILATION
- Input: All optimized prompts
- Skill: #190 (Format and compile)
- Output: Final text file with all prompts

This structure ensures each prompt is crafted with consistent style while being optimized for the specific pun and image generation requirements.

SubAgent #2 - Diagram

Expand Diagram

I'll analyze the "AI Birthday Card Generator" subagent and break it down according to your guidelines.

A) SUBAGENT SUMMARY: 
A specialized image generation agent that takes pre-written puns and image prompts to create a collection of niche-specific birthday cards with integrated text and illustrations in a consistent style.

B) FINAL TASK OUTPUT:
3-5 PNG image files, each 1024x1024 pixels, featuring an AI-generated illustration with integrated text in a birthday card format, with transparent backgrounds where appropriate.

C) SUBAGENT INPUT:
- List of 3-5 finalized pun texts (from Subagent 1)
- List of 3-5 corresponding image prompts (from Subagent 2)
- Optional: Style guidelines for consistent card design

E) SUBAGENT TASK SUMMARY:
For each card (repeated 3-5 times):
Input (pun + prompt) > #222 (Make Image With Text) > #191 (Resize Image if needed) > Final PNG Output

The process uses two primary skills in sequence:
1. #222 - Make Image (Especially With Text)
   - INPUT: Combined prompt including both visual description and pun text
   - OUTPUT: Initial PNG with integrated text

2. #191 - Resize Image (if needed)
   - INPUT: Generated PNG from #222
   - OUTPUT: Resized 1024x1024 PNG file

F) SILOS:
SILO 1: CARD 1
- Input: Pun 1 + Image Prompt 1
- Skill #222: Generate card with text
- Skill #191: Resize if needed
- Output: final_card_1.png

SILO 2: CARD 2
- Input: Pun 2 + Image Prompt 2
- Skill #222: Generate card with text
- Skill #191: Resize if needed
- Output: final_card_2.png

SILO 3: CARD 3
- Input: Pun 3 + Image Prompt 3
- Skill #222: Generate card with text
- Skill #191: Resize if needed
- Output: final_card_3.png

[Additional silos 4-5 follow same pattern if needed]

Each silo operates independently but follows identical steps, ensuring consistent output while allowing for parallel processing if the system supports it. The silos approach ensures each card gets individual attention while maintaining the same quality standards across all outputs.