What Shall We Build Next?

Describe

Describe your task

Refine

Refine the plan

SubAgents

Review all agents

Deploy

Deploy your agent

A) SUBAGENT SUMMARY:  
This subagent (“FunnyRecipeWordsmith”) researches the user’s niche description, then uses that research to generate a set of 3–5 humorous, niche-specific recipes (title, description, ingredients, and steps).

B) FINAL TASK OUTPUT:  
A structured text output (JSON or list) containing 3–5 comedic recipes (the “funny-recipe-text”). Each recipe includes:
• Humorous recipe name  
• Playful, niche-specific description  
• Ingredients list (in a humorous tone)  
• Steps (in a lighthearted, comedic style)

C) SUBAGENT INPUT:  
• 5–200 words describing the niche or audience for which we need humorous recipes.

D) SUBAGENT TASK SUMMARY (Step-by-Step Flow):

1) INPUT → Skill #216 (“Research Topic (Or Keyword) Deeply”)  
     • PURPOSE: Research the niche thoroughly to gather relevant cultural references and comedic hooks.  
     • INPUT: The user’s 5–200 word description of the niche.  
     • OUTPUT: A research summary (1000–3000 characters) with niche-specific insights.

2) Skill #216 OUTPUT → Skill #190 (“Write or rewrite text based on instructions”)  
     • PURPOSE: Transform the research summary into 3–5 comedic recipes.  
     • INPUT:  
       (a) Research summary from Skill #216  
       (b) Instructions to produce comedic recipe titles, descriptions, ingredients, and steps that are niche-specific and humorous.  
     • OUTPUT:  
       • Structured text (JSON or list) of 3–5 humorous recipes, each containing:  
         – Recipe title (humorous)  
         – Description (playful, niche-based)  
         – Ingredients (list in a funny style)  
         – Steps (written humorously)

3) Skill #190 OUTPUT → Subagent Final Output  
     • PURPOSE: Return the comedic recipes in a final structured format suitable for the next subagent to consume (or for direct user output).  
     • OUTPUT: [funny-recipe-text]

E) SILOS (optional breakdown):  
• Silo 1: Niche Research    
  → (Skill #216) Analyze user niche description, produce research insights

• Silo 2: Recipe Creation  
  → (Skill #190) Generate 3–5 humorous recipes based on the research

F) SUBAGENT FINAL OUTPUT:  
[funny-recipe-text]  
A structured text document (e.g., JSON with 3–5 comedic recipes) containing all the required recipe components.

SubAgent #1 - Diagram

Expand Diagram

A) SUBAGENT SUMMARY  
"QuirkyDishIllustrator" is responsible for taking each humorous recipe’s text and creating a whimsical, niche-themed image (PNG) for that recipe, returning 3–5 total images.

B) FINAL TASK OUTPUT  
A list/array of 3–5 transparent PNG image URLs (size 1024×1024) generated from each comedic recipe’s details.

C) SUBAGENT INPUT  
• A structured set of humorous recipes (3–5 recipes), each containing:  
  – Recipe title (funny name)  
  – Playful description  
  – Comedic/niche references, ingredients, steps, etc.

D) SUBAGENT TASK SUMMARY (Step-by-step flow)  
1) Receive the prepared comedic recipe texts (3–5 items).  
2) For each recipe:  
   a) Generate a short text prompt describing the whimsical dish image (Skill #190 - "Write or rewrite text based on instructions"). This uses the recipe’s comedic elements, references, and tone.  
   b) Pass that prompt to Skill #182 - "Create Dalle Image (Transparent Square)" to generate a 1024x1024 transparent PNG.  
3) Collect each generated PNG URL in a single list.  
4) Output that list of 3–5 image URLs as the final subagent result.

E) SILOS

• SILO 1: Prompt Generation for Each Recipe  
  – Input: The comedic recipe text (title, description, cultural references)  
  – Action:  
    1) Skill #190 (“Write or rewrite text based on instructions”)  
       • Prompt: “Using the comedic elements from this recipe, generate a concise but vivid illustration prompt describing a quirky cartoon-style dish that reflects the recipe’s humor and theme.”  
       • Output: A short, comedic illustration prompt

• SILO 2: Image Creation per Recipe  
  – Input: The short comedic illustration prompt from SILO 1  
  – Action:  
    1) Skill #182 ("Create Dalle Image (Transparent Square)")  
       • Output: A 1024×1024 transparent PNG (image URL)

• Final Subagent Output  
  – A list of 3–5 generated image URLs corresponding to each recipe.

SubAgent #2 - Diagram

Expand Diagram

A) SUBAGENT SUMMARY  
"CardLayoutComposer" takes each humorous recipe’s text content plus its associated AI-generated image, then merges them into a single visually appealing recipe card (PNG), repeating for all recipes (3–5).

B) FINAL TASK OUTPUT  
3–5 individual PNG recipe card images (one per recipe), each including the title, description, ingredients, steps, and the corresponding dish illustration.

C) SUBAGENT INPUT  
1) A structured set of 3–5 recipe texts (title, description, ingredients, steps).  
2) A corresponding set of 3–5 image URLs (from Subagent 2).

E) SUBAGENT TASK SUMMARY  
1) Receive the array/list of humorous recipe texts and their matching image URLs.  
2) For each recipe-image pair, feed the text content (title, description, ingredients, steps) plus the dish image URL into skill #222 (Make Image (Especially With Text)) to produce a unified recipe card design.  
3) Collect the resulting 3–5 PNG recipe card URLs as the subagent’s final output.

F) SILOS  
• Silo 1 (“Create Each Recipe Card”):  
  – Input: A single recipe’s text + an image URL.  
  – Action:  
    1) #222 - Make Image (Especially With Text)  
       INPUT: Full text (title, short description, ingredient list, steps) + dish image URL, plus any layout instructions for combining them into a single image.  
       OUTPUT: One PNG card with all text elements and the AI-generated illustration combined.  
  – Repeat this silo 3–5 times (once per recipe).  
• Final Output: The subagent returns a list of 3–5 final PNG URLs.

SubAgent #3 - Diagram

Expand Flow

A) SUBAGENT SUMMARY  
This subagent (“DesignInspirationFinder”) searches for existing funny recipe card designs and provides reference images and a concise text summary for inspiration before creating the final humorous recipe cards.

B) FINAL TASK OUTPUT  
A short text summary (in plain text) describing common design elements, themes, or notable ideas from popular “funny recipe cards for [niche]”, plus a list of up to five relevant image URLs found online that showcase these designs.

C) SUBAGENT INPUT  
• Niche/Topic (e.g., “gamer”, “cat lovers”, “superheroes”, etc.)

E) SUBAGENT TASK SUMMARY

1) Receive the niche/topic as input.  
2) Convert the niche string into a search query, e.g. “funny recipe cards for {niche}”.  
3) Use skill #187 (Find Relevant Images with Brave Search) to gather 5 relevant image URLs.  
   ─ INPUT to #187: The search phrase “funny recipe cards for {niche}.”  
   ─ OUTPUT from #187: A list of 5 image URLs with short descriptions of each.  
4) Use skill #190 (Write or rewrite text based on instructions) to generate a short summary describing the prevalent design elements, styles, or humor trends in those recipe cards.  
   ─ INPUT to #190: The 5 image URLs and instructions: “Summarize the main design ideas and comedic themes found in these images.”  
   ─ OUTPUT from #190: A concise textual overview of observations and inspiration points.

→ SUBAGENT FINAL OUTPUT  
• A short text summary highlighting design insights and comedic styles.  
• A list of up to 5 image URLs (and any brief notes returned by skill #187).

F) SILOS

• SILO 1: Image Search  
  - Single-step usage of #187 with the query “funny recipe cards for {niche}.”

• SILO 2: Create Summary  
  - Single-step usage of #190 to process the returned image links and generate a textual summary.

All steps flow from niche input → image search → comedic design summary → final text + image references.

4 Template & Links

Expand Flow

A) SUBAGENT SUMMARY  
This subagent (“InspirationHarvester”) gathers and reviews existing “funny recipe card” examples (via image searches), then distills insights and design ideas that will ensure our final recipe cards are fresh, unique, and aligned with the niche.

B) FINAL TASK OUTPUT  
A concise, text-based “Inspiration & Design Brief,” summarizing what currently exists in the niche (visually and conceptually) and suggesting 2–4 new angles or design guidelines to differentiate our final recipe cards.

C) SUBAGENT INPUT  
• Niche/topic description (5–200 words)  
• Any relevant style/design preferences (optional)

E) SUBAGENT TASK SUMMARY  
1) Subagent Input → (Skill #187 - “Find Relevant Images with Brave Search”)  
   - Prompt: "funny recipe cards for [niche]"  
   - OUTPUT: 5 image URLs of existing recipe cards or related references.

2) (Skill #177 - “Analyze 4x Images With GPTVision And Return 4x Text”)  
   - INPUT: Up to 4 of the image URLs from step 1 (the 5th can be ignored or used later if needed).  
   - Prompt: "Describe the comedic style, layout, and design elements that make these recipe cards stand out. Summarize any relevant text or features."  
   - OUTPUT: A brief text summary of each of the 4 images (style, humor tone, layout approach).

3) (Skill #190 - “Write or rewrite text based on instructions”)  
   - INPUT:  
     • The summarized image analysis from step 2.  
     • Subagent’s objective: “Generate a short ‘Inspiration & Design Brief’ summarizing key takeaways and 2-4 unique design angles we can adopt.”  
   - OUTPUT: A final text summary (“Inspiration & Design Brief”).

Subagent Final Output: [“Inspiration & Design Brief” as text]

F) SILOS (If needed)  
• Silo 1: Image Collection  
   - #187 retrieves the top 5 sample reference images.

• Silo 2: Image Analysis  
   - #177 takes 4 images, summarizes comedic style, visuals, layout, text presentation.

• Silo 3: Ideation & Brief  
   - #190 compiles the analysis into a cohesive brief, highlighting key takeaways and new, differentiating design ideas.