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 (PromptGenerator) takes an original text input (variable1) and produces three distinct textual prompts describing visual scenes that can be used to generate images representing that original text.

B) FINAL TASK OUTPUT:  
A single text output containing three separate image-generation prompts (e.g., “1)… 2)… 3)…”).

C) SUBAGENT INPUT:  
• The original text input (variable1) – a line of text, script, or paragraph.  
• An instruction telling the model to create three distinct image prompts (e.g., “Generate three distinct image prompts”).

E) SUBAGENT TASK SUMMARY:  
1. Subagent receives the original text input (variable1) and the instruction to “Generate three distinct image prompts.”  
2. Subagent calls Skill #190 (“Write or rewrite text based on instructions”) with the following combined prompt:  
   - “Using the text below, create three distinct, creative, visually rich image prompts that represent the essence or message of the text in unique ways. Output them clearly labeled as 1), 2), 3).”  
   - Insert variable1 content below that instruction.  
3. Skill #190 outputs a single textual response that includes all three prompts.  
4. Subagent returns that text as the final result ([3-image-prompts]).

F) SILOS (if any):  
There are no sub-silos required here. The subagent performs only one sequential task—calling Skill #190 once and returning its result.

SubAgent #1 - Diagram

Expand Diagram

A) SUBAGENT SUMMARY  
This subagent takes three text prompts, creates one AI-generated PNG image per prompt, and finally returns a concise list of the three image URLs.

B) FINAL TASK OUTPUT  
A single text response containing three PNG image URLs (an array or list), each image corresponding to one of the three prompts.

C) SUBAGENT INPUT  
• [3-image-prompts]: An array of three distinct text prompts describing what image to generate.

D) SUBAGENT TASK SUMMARY

1) Receive the three prompts as [3-image-prompts].  
2) For each prompt (Prompt #1, #2, #3):  
   – Call Skill #222 (“Make Image (Especially With Text)”) and generate a PNG image URL.  
3) Collect the resulting 3 PNG URLs.  
4) Call Skill #223 (“Powerful LLM Prompt-to-Text Response”) to format these 3 URLs into a simple final list.  
5) Return the final text output containing the list of 3 PNG links (the subagent’s final output).

E) SILOS

• Silo 1: Generate Images (Repeated Steps for Each of the 3 Prompts)  
  – Input: Single prompt from [3-image-prompts]  
  – Action: Use Skill #222 (Make Image) → Output: PNG URL

• Silo 2: Collate and Return URLs  
  – Input: 3 PNG URLs  
  – Action: Use Skill #223 (Powerful LLM Prompt-to-Text Response) to produce a final text listing of the three URLs  
  – Output: [3-generated-images] (a text list or JSON array of the final PNG URLs)

SubAgent #2 - Diagram

Expand Diagram

A) SUBAGENT SUMMARY  
This subagent (VideoAnimator) takes three text prompts (originally used to generate images) and creates three separate short (~5-second) animated MP4 videos from those prompts, then returns the three final MP4 URLs in a concise list.

B) FINAL TASK OUTPUT  
[3-animated-videos] → A list/array of exactly three MP4 URLs, each corresponding to a short animated video.

C) SUBAGENT INPUT  
• [3-image-prompts] → The three text prompts used to describe the visual scenes.  
• [3-generated-images] → The three image URLs (though not directly used for animation, provided as context if needed).

E) SUBAGENT TASK SUMMARY  
1) Input:  
   – Subagent receives [3-image-prompts] and [3-generated-images].  
2) Generate the 3 short animated videos:  
   – For each prompt in [3-image-prompts], call skill #169 (Create A 5 Second Animated Video From Text Prompt).  
   – Each call to skill #169 will accept one text prompt and output one MP4 URL.  
3) Compile the three MP4 URLs into a single list:  
   – After all three videos are created, call skill #223 (Powerful LLM Prompt-to-Text Response) to format those three MP4 URLs into a simple, labeled list or array.  
4) Output:  
   – Return the final [3-animated-videos] (three MP4 URLs).

F) SILOS

• Silo 1: Generate Three Videos  
   – Action 1 (repeated ×3):  
     Skill #169 → INPUT: single text prompt from [3-image-prompts]; OUTPUT: MP4 video URL.  
   – Collect the three MP4 URLs.

• Silo 2: Compile URLs into Final Output  
   – Action 2:  
     Skill #223 → INPUT: the three MP4 URLs; OUTPUT: a formatted text response listing these three MP4 URLs neatly (i.e., “[3-animated-videos]”).

End Result: [3-animated-videos] – three final MP4 URLs, each ~5 seconds long.

SubAgent #3 - Diagram

Expand Flow

A) SUBAGENT SUMMARY  
The “ResultAssembler” subagent will take the three prompts, the three generated image URLs, and the three animated video URLs, then format them into a single JSON object so that each item includes its prompt, its image, and its video.

B) FINAL TASK OUTPUT  
A JSON object containing three entries (each containing one prompt, the corresponding image URL, and the corresponding video URL). Example structure:  
{
  "results": [
    {
      "prompt": "Prompt #1",
      "image": "ImageURL #1",
      "video": "VideoURL #1"
    },
    ...
  ]
}

C) SUBAGENT INPUT  
• [3-image-prompts] (the three text prompts)  
• [3-generated-images] (the three image URLs)  
• [3-animated-videos] (the three MP4 video URLs)

E) SUBAGENT TASK SUMMARY  
1) Receive the three prompts, three image URLs, and three video URLs as input.  
2) Use Skill #190 (Write or rewrite text based on instructions) to arrange these nine total data points into a single JSON object, ensuring the structure includes (prompt, image, video) for each of the three items.  
3) Return the JSON object text as the final subagent output.

F) SILOS

• Silo 1 – Formatting and Assembling JSON:  
  – Input: [3-image-prompts], [3-generated-images], [3-animated-videos].  
  – Action:  
    • Skill #190 – Write or rewrite text based on instructions.  
      – Description/Purpose: Uses the LLM to create the JSON output with a natural JSON structure mapping prompt-to-image-to-video for each of the three sets.  
      – Output: [video-urls-json] (the final JSON text).  
  – Output: The final structured JSON object.

4 Template & Links

Expand Flow

5 Template & Links

Expand Flow

Below is a template showing how you might lay out a “Subagent #6” in the same style as the others, even though no Subagent #6 was explicitly defined. This example simply demonstrates the requested format. If in your larger workflow you truly have no need for a “Subagent #6,” you can safely ignore this. But if you do decide a sixth subagent is needed (for example, to handle video post-processing, thumbnail generation, advanced logging, or any other specialized purpose), you can follow this style when defining it.

──────────────────────────────────────────────────────────
A) SUBAGENT SUMMARY
──────────────────────────────────────────────────────────
This subagent (labeled here as “Subagent #6”) receives inputs from earlier subagents (e.g., the three MP4 videos already generated) and performs any additional specialized tasks required for the workflow. It then outputs its final result, which is narrowly defined for its unique function.

──────────────────────────────────────────────────────────
B) FINAL TASK OUTPUT
──────────────────────────────────────────────────────────
• The final files or data that this Subagent #6 produces (for example, three individually post-processed MP4 videos, or a single combined MP4, or a JSON object, etc.).  
• Must be specified in a precise technical format (e.g., “Three MP4 URLs,” “One combined MP4,” “JSON text,” etc.).

──────────────────────────────────────────────────────────
C) SUBAGENT INPUT
──────────────────────────────────────────────────────────
• The necessary data for Subagent #6 to do its job.  
• For instance:  
  1) The three MP4 URLs from the previous subagent(s).  
  2) Any relevant text prompts, metadata, or instructions required to carry out its specialized task.

──────────────────────────────────────────────────────────
E) SUBAGENT TASK SUMMARY
──────────────────────────────────────────────────────────
Below is the step-by-step flow of skills you might chain together. This is just one illustration; replace or revise steps and skill numbers as appropriate for the actual specialized job of Subagent #6.

1) SUBAGENT INPUT(S)  
   └─ (e.g., receives three MP4 URLs and a request to “enhance or augment each video.”)

2) [Optional] Use Skill #190 (Write or rewrite text based on instructions)  
   └─ PURPOSE: If you need to generate new textual instructions or a short piece of copy to overlay on each video, you can do that here.  
   └─ INPUT: A prompt describing the kind of text or instructions needed.  
   └─ OUTPUT: Text snippet(s) to be used in the next step.

3) [Optional] Use Skill #199 (Add Images & Videos On Top Of Existing MP4)  
   └─ PURPOSE: For instance, if you are overlaying text or images on the existing MP4 to create final, post-processed clips.  
   └─ INPUT: Each MP4 URL + the textual instructions or assets.  
   └─ OUTPUT: Up to three new MP4 URLs (one for each augmented video).

4) [Optional] Use Skill #173 (Concatena Videos)  
   └─ PURPOSE: If Subagent #6 needs to merge multiple videos into one final output.  
   └─ INPUT: A list of 2–10 MP4 videos.  
   └─ OUTPUT: Single MP4 URL with all videos concatenated.

5) [Optional] Use Skill #190 or #223 again to format or finalize the output.  
   └─ PURPOSE: Return the final link(s) in a well-structured format (e.g., JSON).  
   └─ INPUT: The final MP4 URL(s) and any relevant text to shape the output.  
   └─ OUTPUT: A textual response containing the final data, URIs, or JSON result.

6) SUBAGENT FINAL OUTPUT  
   └─ The subagent’s end-product (e.g., final processed MP4(s), or a JSON object listing them).

──────────────────────────────────────────────────────────
F) SILOS
──────────────────────────────────────────────────────────
If Subagent #6 performs multiple mini-processes in parallel (or distinctly partitioned major steps), you can break them into silos:

• SILO 1: “Text Generation”  
  (Example: Step where Subagent #6 writes or refines any required text/captions.)

• SILO 2: “Video Enhancements”  
  (Example: Steps adding elements onto each MP4, overlaying images/text, or merging them.)

• SILO 3: “Output Formatting”  
  (Example: Step returning final data as JSON or some other structure.)

──────────────────────────────────────────────────────────

This example shows how you would define a Subagent #6 with clear inputs, outputs, steps, and skill usage. Adapt it to the actual purpose and skill chain that your newly created subagent must fulfill. If in your workflow there truly is no Subagent #6, you can disregard this skeleton. Otherwise, fill in the details as needed to ensure smooth handoff between any previous subagent’s output and your final required deliverable.