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 (“BusinessSearch”) obtains up-to-date business information (e.g., name, address, phone, website) by querying the web as if performing a Google search, then returns a text-based list of businesses.

B) FINAL TASK OUTPUT  
Raw text-based answer (string) containing multiple business listings, each with basic identifying details (City, Business Name, Phone, Street Address, ZIP, Website).

C) SUBAGENT INPUT  
• businessType (e.g., “plumbers”)  
• location (e.g., “Dallas, Texas”)

D) FINAL TASK OUTPUT (Restated More Technically)  
[business-listings-text]  
A plain text string containing the results of the search query. Each line or section should represent a distinct business record, listing name, phone, address, ZIP, website (as available).

E) SUBAGENT TASK SUMMARY (Step-by-Step Flow)

1) (Optional) Generate Oracle Search Prompt  
   – Use skill #190 - “Write or rewrite text based on instructions”  
     INPUT: A small piece of text, for example:  
       “The user wants a list of [businessType] in [location], including name, phone, street, city, ZIP, website. Format the request as a succinct search query for the Oracle.”  
     OUTPUT: A short text prompt to be fed into the Oracle (e.g., “List [businessType] in [location], including the name, phone, address, ZIP, website.”)

2) Query the Oracle (with Live Web Access)  
   – Use skill #224 - “Oracle Ask A Question”  
     INPUT: The query text from step #1 (or directly from subagent input if rewriting is skipped).  
     OUTPUT: A text-based answer that includes multiple business listings matching the user’s query.

Subagent Final Output = [business-listings-text]  
(The raw text from the Oracle describing the businesses.)

F) SILOS

• SILO 1: Prompt Preparation  
  – Takes (businessType, location)  
  – Uses skill #190 to craft a succinct Oracle query  
  – Output: “List plumber businesses in Dallas, TX, including name, phone, street, city, ZIP, and website.”

• SILO 2: Oracle Live Search  
  – Takes the prompt from SILO 1  
  – Uses skill #224 to get live data  
  – Output: multi-line text with business listings

After SILO 2 completes, the subagent returns the raw text as [business-listings-text].

SubAgent #1 - Diagram

Expand Diagram

A) SUBAGENT SUMMARY  
This subagent (“DataFormatter”) takes raw/unstructured text containing business listings and converts it into a clean, standardized JSON array of business objects.

B) FINAL TASK OUTPUT  
A single JSON-formatted string that contains an array of objects with the following fields for each business:  
• city  
• businessName  
• phoneNumber  
• streetAddress  
• zipCode  
• websiteUrl

C) SUBAGENT INPUT  
• The raw text listing of businesses (from Subagent #1).  
• Instructions describing how to parse that text into JSON objects containing (city, businessName, phoneNumber, streetAddress, zipCode, websiteUrl).

E) SUBAGENT TASK SUMMARY  
1) Subagent receives the unstructured text output (“business-listings-text”) from Subagent #1.  
2) Subagent calls Skill #223 (“Powerful LLM Prompt-to-Text Response”) with a detailed prompt:  
   • Provide the unstructured text.  
   • Clearly request JSON output in the specified schema (city, businessName, phoneNumber, streetAddress, zipCode, websiteUrl) for each business listing found in that text.  
3) The LLM returns JSON-formatted text containing an array of objects satisfying the required fields.  
4) Output the final JSON text as “structured-business-data.”

F) SILOS  
Not strictly required. This subagent flow is handled in one main step:  
• Silo 1: Parse and structure data → (Skill #223)  
• Silo 2: Return final JSON

SubAgent #2 - Diagram

Expand Diagram

A) SUBAGENT SUMMARY  
The SpreadsheetGenerator subagent takes structured JSON data containing business information and generates a single CSV-formatted text output that includes Crescent Technical branding and can be opened in Excel.

B) FINAL TASK OUTPUT  
A multi-line CSV text string containing:
• Columns: City, Business Name, Phone Number, Street Address, ZIP Code, and Website URL  
• A mention of "Crescent Technical" (e.g., as a header note or extra row)  
• Ready to be saved as .csv or imported into Excel

C) SUBAGENT INPUT  
• The structured JSON array from Subagent #2 (each entry containing fields for “city,” “businessName,” “phoneNumber,” “streetAddress,” “zipCode,” and “websiteUrl”)  
• Instruction to insert “Crescent Technical” branding into the output

E) SUBAGENT TASK SUMMARY  
1) Receive the structured JSON array from Subagent #2 along with the requirement to include “Crescent Technical.”  
2) Use Skill #190 (Write or rewrite text based on instructions) to transform the JSON data into CSV text. The instruction prompt to Skill #190 will specify:  
   • Create column headers: City, Business Name, Phone Number, Street Address, ZIP Code, Website URL  
   • Generate rows for each JSON object’s data in order.  
   • Insert mention of “Crescent Technical” (for branding).  
3) Output the final multi-line CSV string.

F) SILOS  
• Silo 1: Transform JSON to CSV  
   - Input: JSON and instructions  
   - Action (Skill #190): Produce CSV text rows, include branding  
   - Output: Final CSV text string (Excel-ready).

SubAgent #3 - Diagram

Expand Flow

A) SUBAGENT SUMMARY  
This (hypothetical) Subagent #4 is responsible for taking the CSV-formatted output from the SpreadsheetGenerator (Subagent #3) and delivering it as a final Excel (.xlsx) file containing the requested business listings with “Crescent Technical” branding.

B) FINAL TASK OUTPUT  
A single .xlsx file (Excel workbook) that lists all matching businesses in the requested location. Each row contains columns for City, Business Name, Phone Number, Street Address, ZIP Code, and Website URL, plus a mention or small footer referencing “Crescent Technical.”

C) SUBAGENT INPUT  
• CSV text or similar spreadsheet-compatible data structure from Subagent #3.  
• (Optionally) Any user instructions regarding styling, such as how the “Crescent Technical” branding should appear (e.g., as a header row, footer note, or text label).

D) SUBAGENT TASK SUMMARY  
Below is the step-by-step outline of how Subagent #4 would transform a CSV string into a final .xlsx file and include branding. Note that there is no direct skill to create .xlsx files in the current skill set; therefore the approach uses a workaround. It is assumed an LLM-based rewrite/bundling approach can embed the CSV data in a base64-encoded .xlsx file for downloading (or, as an alternative, simply return CSV that the user can rename in Excel).

1) Subagent receives “CSV-formatted text” from Subagent #3.  
2) Use Skill #190 (“Write or rewrite text based on instructions”) to:  
   • Generate either a textual placeholder for an .xlsx file, or produce instructions/base64 that can be turned into an .xlsx file.  
   • Add “Crescent Technical” branding (e.g., as a simple textual mention in the first row or last row).  
   • If no actual .xlsx conversion is possible within the skill set, the final text output will remain CSV (with branding inserted) and a user instruction to “Open or save this file as .xlsx.”  
3) The final step is to return the newly created text (CSV or base64-converted to .xlsx) as the subagent’s output.

E) SILOS  
Since this is a fairly straightforward subagent with one primary objective—converting or packaging CSV data into an actual downloadable Excel file—the sub-tasks can be grouped in a single silo:

• SILO 1:  
   - Input: CSV text (including columns for City, Business Name, Phone Number, Street Address, ZIP Code, Website URL).  
   - Action Step 1 (Skill #190): Append “Crescent Technical” branding, rewrite as needed (either CSV or a pseudo-.xlsx/base64).  
   - Output: Final subagent output (usable file content).

––––––––––––––––––

In practice, if the platform or environment supports generating real .xlsx files, an actual file-conversion step would occur here. With the current available skills, the best approximation is to generate a CSV (or base64 version) that can be downloaded as “filename.xlsx” and opened in Excel.

4 Template & Links

Expand Flow

A) SUBAGENT SUMMARY  
Subagent #5 is responsible for taking the CSV-like output from prior steps (e.g., Subagent #3), branding it with “Crescent Technical,” and producing a final Excel file (.xlsx) that contains the user-requested columns for City, Business Name, Phone Number, Street Address, ZIP Code, and Website URL.

B) FINAL TASK OUTPUT  
A single .xlsx file (Excel spreadsheet) which includes:  
• Headers: City, Business Name, Phone Number, Street Address, ZIP Code, Website URL  
• All business data rows  
• Prominent mention or branding of “Crescent Technical” (e.g., in the top row or header area)

C) SUBAGENT INPUT  
• The final CSV text output from Subagent #3 (containing columns for City, Business Name, Phone Number, Street Address, ZIP Code, Website URL)  
• Any user instruction about how/where to place the “Crescent Technical” branding (e.g., “Please include my brand name in the first row’s first cell”).

E) SUBAGENT TASK SUMMARY  
Below is the breakdown of how Subagent #5 will produce this final .xlsx file. Since there is no direct “convert CSV to XLSX” skill in the available skill list, we will use text-based transformations to approximate or generate an XLSX data structure (often either returned in Base64 or as a pseudo-file). Then we will provide that final output so the user can receive and download it.

1) Receive CSV Data & Branding Instructions  
   • INPUT:  
     – CSV text from Subagent #3 (containing header row + data rows).  
     – Branding instructions from user.  
   • ACTION: Store these inputs locally to begin file creation.

2) Parse the CSV Data (optional)  
   • SKILL USED: #223 - Powerful LLM Prompt-to-Text Response  
   • PURPOSE:  
     – We can use an LLM to parse the CSV text into an internal structured representation (like a list of row objects).  
   • INPUT NEEDED:  
     – The CSV text.  
     – A prompt instructing the LLM to parse the CSV into structured objects.  
   • OUTPUT:  
     – Structured JSON or an internal text representation of the rows/columns.

3) Generate XLSX-Format or Equivalent Base64 (Approximation)  
   • SKILL USED: #190 - Write or rewrite text based on instructions  
   • PURPOSE:  
     – Convert the structured data into an Excel-compatible file representation, possibly a Base64-encoded .xlsx, along with the “Crescent Technical” branding in the first row (or whichever area the user desires).  
   • INPUT NEEDED:  
     – The structured data from step (2).  
     – Instructions on how to incorporate branding text (e.g., “Crescent Technical” as a header or footnote).  
   • OUTPUT:  
     – A text-based output that represents the content of our final .xlsx. This is typically provided as a downloadable file:  
       For example: “data:application/vnd.openxmlformats-officedocument.spreadsheetml.sheet;base64, <base64>”  
     – Alternatively, if the environment does not support direct .xlsx creation, produce instructions or a rendered mock XLSX that can be downloaded by the user and opened in Excel.

4) Return Final File  
   • Subagent #5 Final Output: The .xlsx file (or .xlsx-equivalent) is returned, ready to be delivered to the user.

F) SILOS (Optional)  
• Silo 1: Input & Preparation  
   – Gather CSV text from Subagent #3, get branding instructions.  
• Silo 2: Structuring & Data Handling  
   – Parse CSV to an internal format.  
• Silo 3: XLSX Synthesis  
   – Produce or encode final spreadsheet with the “Crescent Technical” banner.  
• Silo 4: Delivery  
   – Combine into final .xlsx file output.

This completes Subagent #5’s workflow for generating and returning the final Excel (.xlsx) file with branding.

5 Template & Links

Expand Flow

A) SUBAGENT SUMMARY  
Subagent #6 is responsible for taking the user’s request (business type + location), searching the web for matching businesses, structuring that data, and producing a final Excel-ready file with the requested columns and Crescent Technical branding.

B) FINAL TASK OUTPUT  
A single Excel-ready (.csv or .xlsx) file (which can be opened in Excel) containing rows of businesses with the following columns:  
• City  
• Business Name  
• Phone Number  
• Street Address  
• ZIP Code  
• Website URL  
Plus “Crescent Technical” branding somewhere in the file (e.g., as a header or footer note).

C) SUBAGENT INPUT  
• The user’s search parameters:  
  1. The business type (e.g., “Plumbers”)  
  2. The location (e.g., “Dallas, Texas”)

E) SUBAGENT TASK SUMMARY  
Below is the step-by-step flow for Subagent #6. The chain of skills is shown in order:

1) Receive user request.  
   - Input example: “Plumbers in Dallas, Texas.”

2) Search for up-to-date business listings.  
   - SKILL USED: #224 - Oracle Ask A Question  
   - INPUT: A prompt such as “List 10 plumbers in Dallas, Texas including name, phone, street, city, ZIP, website.”  
   - OUTPUT: Text-based answer with multiple business listings (unstructured).

3) Parse the unstructured listings into structured JSON data.  
   - SKILL USED: #223 - Powerful LLM Prompt-to-Text Response  
   - INPUT:  
       • The text from step (2).  
       • Instructions to parse it into a JSON array, each object containing:  
         {  
           “city”: “…”,  
           “businessName”: “…”,  
           “phoneNumber”: “…”,  
           “streetAddress”: “…”,  
           “zipCode”: “…”,  
           “websiteUrl”: “…”  
         }  
   - OUTPUT: A JSON array of objects (structured-business-data).

4) Generate a “CSV-like” text string for Excel, inserting Crescent Technical branding.  
   - SKILL USED: #190 - Write or rewrite text based on instructions  
   - INPUT:  
       • The structured JSON data from step (3).  
       • Instructions to produce comma-separated rows with columns:  
         “City,Business Name,Phone Number,Street Address,ZIP Code,Website URL”  
         Plus a row or note stating “Crescent Technical” for branding.  
   - OUTPUT: A multiline CSV string that can be opened in Excel as “.csv” or saved as “.xlsx.”

5) Final output returned by subagent.  
   - The CSV (or .xlsx) file, ready for download or further use.

F) SILOS  
Since Subagent #6 does a straightforward linear flow, we can conceptually break it into three silos:

• Silo 1 (Data Gathering):  
  - Use #224 (Oracle Ask A Question) to find current business listings.

• Silo 2 (Data Structuring):  
  - Use #223 (Powerful LLM Prompt-to-Text Response) to parse the unstructured listings into well-formed JSON.

• Silo 3 (Spreadsheet Creation):  
  - Use #190 (Write or rewrite text based on instructions) to format JSON data as a CSV-like output with “Crescent Technical” branding.

This completes the workflow for Subagent #6.