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, "Web Search & Discovery," takes a combined search string (niche, subniche, location, special notes) then performs in-depth online research to find URLs or references to small, recently launched businesses that meet the specified criteria (e.g., low domain authority, small social media presence, not sponsored, etc.). It returns a textual research summary listing potential business URLs or references.

B) FINAL TASK OUTPUT  
A single text block (the “[web-search-discovery-list]”) containing a research summary with URLs or references to potential businesses that match the query.

C) SUBAGENT INPUT  
• Niche (e.g. “Dentist”)  
• Subniche (e.g. “Orthodontist”)  
• City (e.g. “Austin”)  
• State (e.g. “Texas”)  
• Country (e.g. “USA”)  
• Special notes (e.g. “newly launched,” “low domain authority,” “exclude sponsored listings,” “small social presence”)

E) SUBAGENT TASK SUMMARY  
1) Generate a refined search-query text (using Skill #190: “Write or rewrite text based on instructions”)  
   • INPUT: Subagent input (niche, subniche, location, special notes)  
   • ACTION: Craft a consolidated text prompt/search query that includes all user-provided parameters (e.g., “Dentist orthodontist Austin Texas USA newly launched no domain authority small social media presence exclude sponsored listings”).  
   • OUTPUT: Single “refined search-query” string.

2) Execute deep research with the refined query (using Skill #216: “Research Topic (Or Keyword) Deeply”)  
   • INPUT: The “refined search-query” string from step 1.  
   • ACTION: Perform multiple web and search-engine checks (Google, Bing, etc.) to find smaller, newly launched businesses that match the criteria in the query.  
   • OUTPUT: A 1000–3000 character textual research summary, potentially containing references, URLs, or short snippet results.

3) Filter and finalize results (using Skill #223: “Powerful LLM Prompt-to-Text Response”)  
   • INPUT: The raw textual research from step 2, plus instructions to exclude big brands, large companies, or sponsored listings, and keep only smaller, recently launched websites with low domain authority and minimal social presence.  
   • ACTION: Parse and refine the textual summary to ensure only websites or references that meet the requested criteria are included.  
   • OUTPUT: The final “[web-search-discovery-list]” — a textual research summary enumerating businesses/URLs that satisfy the user’s parameters.

F) SILOS

• Silo 1: Query Generation  
  - Step: Use Skill #190 to assemble a single, nuanced search phrase that includes all location parameters, niche/subniche, and any special notes.

• Silo 2: Deep Web Research  
  - Step: Pass the refined query from Silo 1 to Skill #216 to gather in-depth text referencing potential business leads.  
  - Output: Detailed text summary containing discovered URLs.

• Silo 3: Results Filtering & Final Summary  
  - Step: Use Skill #223 to process the text from Silo 2. Remove big/sponsored/larger entities, keep only smaller, newly launched businesses with low domain authority. Output the final text block “[web-search-discovery-list].”

SubAgent #1 - Diagram

Expand Diagram

A) SUBAGENT SUMMARY  
"Domain & Contact Info Extraction" focuses on taking a list of potential business URLs, determining whether each site meets domain-related criteria (low authority, recent launch), and then extracting contact information such as main contact name, phone, email, and any social links.

B) FINAL TASK OUTPUT  
The final output is a structured dataset (array/object) labeled [domain-contact-info-extraction]. For each valid URL, it will include:  
• URL  
• Approximate Launch Date  
• Domain Authority (if computed/known or approximate)  
• Main Contact Person Name  
• Phone Number  
• Email Address  
• Additional Contact/Social Links

C) SUBAGENT INPUT  
1. A list/array of URLs discovered by Subagent 1 (e.g., [web-search-discovery-list])  
2. Instructions on which data to extract and how to verify (domain authority, launch date, contact details, etc.)

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

1) Receive the array of candidate URLs (plus instructions).  
   – Input: [web-search-discovery-list] (an array/list of URLs)  
   – No specific skill used here—just the subagent’s starting input.

2) For each URL in the list, gather domain-level information:  
   a) Use SKILL #224 - Oracle Ask A Question  
      – Input: A query about domain authority, recent launch indicators, or any domain data needed (e.g., “Please gather approximate domain authority, spam score, date launched, any relevant business-size indicators for [URL].”)  
      – Output: A text response with best estimations of domain authority, spam score, approximate launch date, etc.

3) Decide whether the site meets the criteria (low domain authority < 30, recently launched, small business, etc.).  
   – No direct skill needed; the subagent will parse the Oracle’s response to see if the site qualifies.  
   – If it does not meet the criteria, skip to next URL.  
   – If it does, proceed.

4) Extract structured data from the qualified URL to get contact details:  
   a) Use SKILL #226 - Extract Structured Data From 1x URL  
      – Input: The specific URL and instructions, such as “Extract any contact person’s name, phone number, email addresses, and any social media links.”  
      – Output: Structured text (e.g., JSON or structured key-value) containing the site’s contact details.

5) Optional validity checks or clarifications (if needed):  
   a) Use SKILL #224 - Oracle Ask A Question  
      – Input: “Verify if [extracted email] is functioning, or if phone [extracted phone] is correct.”  
      – Output: Additional info verifying email/phone, or returning “no data found.”  
   – The subagent then merges any valid verification data into its final structured record for that URL.

6) Build the final record for each valid URL. This record includes:  
   – URL  
   – Approximate Launch Date (from Oracle’s data)  
   – Domain Authority or relevant domain info  
   – Main Contact Person Name  
   – Email Address (or addresses)  
   – Phone Number  
   – Additional Contact/Social Links

7) Continue until all URLs are processed.

8) Return the final aggregated list of records as the subagent’s output: [domain-contact-info-extraction].

F) SILOS  
• Silo 1: “Domain Authority & Launch Data Gathering”  
  – Input per URL → SKILL #224 (Oracle) to gather domain authority, approximate launch date, spam score, possible business size.  
  – Output: Text data describing domain authority, spam score, business size, etc.

• Silo 2: “Contact Info Extraction”  
  – Input: Qualified URL (meets small/new domain condition)  
  – SKILL #226 (Extract Structured Data From 1x URL) to extract contact name, email, phone, social links.  
  – Output: Structured contact info.

• Silo 3: “Validity, Consolidation, Final Output”  
  – Optionally re-ask Oracle (#224) for verification of email/phone if needed.  
  – Consolidate data into a single structured record per URL.  
  – Aggregate into subagent’s final output array [domain-contact-info-extraction].

SubAgent #2 - Diagram

Expand Diagram

A) SUBAGENT SUMMARY  
This subagent (“Output Compiler”) takes the final, validated business records from Subagent 2, sorts and filters them as needed, and then outputs two things: (1) a text block representing a CSV file with specific columns, and (2) a snippet listing all email addresses. It concludes by returning “Mission Complete.”

B) FINAL TASK OUTPUT  
1) A text block containing CSV data with columns:  
   • Website URL  
   • Main Contact Person Name  
   • Email Address  
   • Phone Number  
   • Additional Contact/Social Links

2) A separate text snippet (line- or comma-separated) containing only the email addresses.  
3) A “Mission Complete” message appended at the end.

C) SUBAGENT INPUT  
• A structured list/array of business records (e.g., from Subagent 2), with each record containing:  
  - Website URL  
  - Contact Person Name(s)  
  - Email(s)  
  - Phone Number(s)  
  - Additional Contact/Social Links  
  - Company age or launch date  
  - Any other relevant metadata for filtering

E) SUBAGENT TASK SUMMARY  
1) Receive the structured data from Subagent 2.  
2) Use skill #223 (Powerful LLM Prompt-to-Text Response) to:  
   • Sort the list by company age (newest first, or per final instructions).  
   • Discard any larger companies if discovered (based on any “large” indicator in the data).  
   • Format each remaining record into CSV lines under columns:  
     [Website URL, Main Contact Person Name, Email Address, Phone Number, Additional Contact/Social Links].  
   • Collect all email addresses and produce a separate snippet (line- or comma-separated).  
   • Add “Mission Complete” text at the end of the overall output.  
3) Return the final output from skill #223 as the subagent’s output.

F) SILOS  
• SILO 1: “Data Preparation & Sorting”  
  - Validate that Subagent 2’s data arrives in a structured format.  
  - Insert instructions into skill #223 to sort data by company age and remove large companies.

• SILO 2: “CSV Compilation & Email List Extraction”  
  - Within the same skill #223 call, instruct the LLM to convert the sorted data into CSV lines and extract the email addresses into a separate snippet.

• SILO 3: “Finalize Output”  
  - Instruct skill #223 to append “Mission Complete” and return the CSV text block plus the email list snippet.

SubAgent #3 - Diagram

Expand Flow

A) SUBAGENT SUMMARY  
This subagent is responsible for taking each prospective business record (including URL, contact info, etc.) and performing a rigorous validation check—verifying that the business is recently launched (≤3 years), has low or no domain authority, is small (limited social presence and not a large corporate entity), and that all gathered contact details (website URL, email, phone) are accurate and active.

B) FINAL TASK OUTPUT  
A filtered, verified list of business records that pass all the criteria (recently launched, low DA, valid contact details, small social presence). The subagent returns that verified dataset as a structured text/object that can be passed on for final compilation (e.g. to the Output Compiler subagent).

C) SUBAGENT INPUT  
• A collection of business records (URLs plus preliminary contact details from previous subagents). Each record should minimally include:  
  1) Website URL  
  2) Potential main contact person name (if found)  
  3) Potential email address  
  4) Potential phone number  
  5) Additional notes (e.g., possible social links)

E) SUBAGENT TASK SUMMARY

1) For each business record in the subagent input set, perform the following sequence of skills/actions:

(a) Determine Overall Domain Authority & Business Scale  
      • Use “#224 - Oracle Ask A Question” with a query such as:  
        “Please check <DOMAIN> on SEMrush (or similar) to retrieve:  
         1) Domain Authority or Domain Rating (approximate OK).  
         2) If the business or brand is large or small.  
         3) Any major indicators that it’s a well-established (older than 3 years) or large corporation.”  
      • Parse the Oracle’s answer to see if DA < 30 and business is not a large corporate entity.  
      • If DA ≥ 30 or obviously a large company, exclude this record.

(b) Verify Recent Launch (≤3 years)  
      • Again use “#224 - Oracle Ask A Question” with text like:  
        “Has <BUSINESS OR DOMAIN> been launched in the last 3 years? Check official databases (government sites) or any press/news coverage.”  
      • If the Oracle’s answer indicates an older launch or no data that suggests it is new, exclude this record.

(c) Verify Contact Info Is Valid / Active  
      • Phone Number Check:  
         “#224 - Oracle Ask A Question”  
         e.g. “Does phone number <XYZ> appear to be valid and currently in use by <BUSINESS NAME>?”  
         If obviously invalid, exclude.  
      • Email Address Check:  
         “#224 - Oracle Ask A Question”  
         “Does email <XYZ> appear to be valid, e.g. publicly referenced or recognized for <DOMAIN>?”  
         If invalid, exclude.

(d) Check Social Presence Is Small  
      • “#224 - Oracle Ask A Question”  
         “Evaluate <BUSINESS NAME or DOMAIN> on social media. Is the social following minimal (small or rarely used accounts)? Is there any mention of very large brand presence?”  
      • If the brand has an obviously big social following or presence, exclude.

(e) Output Verified Record  
      • If a record passes all checks, the subagent outputs it in a structured format (e.g. JSON-like or table-like text) that includes official “verified = true” status.  
      • If it fails, it’s excluded or flagged as “verified = false.”

2) After processing every record, the subagent returns the final verified set.

F) SILOS

• Silo 1: Domain Authority & Scale Check  
  – Input: Single business record → #224 Oracle question about domain authority & brand size → Output: DA value & brand-size notes  
  – Logic: If DA < 30 and brand not large, proceed. Else exclude.

• Silo 2: Recent Launch Verification  
  – Input: The same record → #224 Oracle question about business launch date → Output: Launch date/approx.  
  – Logic: If ≤3 years, proceed. Else exclude.

• Silo 3: Contact Data Validation  
  – Input: The same record → #224 Oracle question for phone + #224 Oracle question for email → Output: Valid/invalid status for phone/email.  
  – Logic: If either is invalid, exclude.

• Silo 4: Social Presence Check  
  – Input: The same record → #224 Oracle question about brand’s social presence → Output: Small or large presence.  
  – Logic: Exclude if presence is large.

• Silo 5: Final Collation  
  – For all records that pass the above silos, return them in a verified data structure.  
  – This final verified output is then ready for the next stage (e.g. the output-compiler subagent).

4 Template & Links

Expand Flow

A) SUBAGENT SUMMARY  
This subagent re-checks the raw list of discovered businesses and their contact details to ensure they genuinely match the “recently launched, low domain authority, almost no web presence, small social following” criteria. It also verifies that each website, email address, phone number, and contact person are legitimate.

B) FINAL TASK OUTPUT  
A filtered, validated dataset (the same structured format as the input but now confirmed to meet the specified criteria or excluded otherwise), ready for the next step in the workflow. Specifically, this subagent returns an updated list of businesses that satisfy:  
• Launched in the past 3 years  
• Low domain authority (<30) and minimal backlinks  
• Very low traffic / “hard to find”  
• Website functional and actively owned by the listed contact person(s)  
• Contact info (email, phone, etc.) is confirmed to be valid

C) SUBAGENT INPUT  
• A structured list or array of businesses, each with:  
  – Website URL  
  – (Optional) Preliminary domain authority info  
  – Main Contact Name  
  – Email Address  
  – Phone Number  
  – Any Additional Links/Social Profiles  
  – Any preliminary note(s) about recent launch date or domain age

D) SUBAGENT OUTPUT  
• A validated/filtered list in the same structured format, but now each record is either:  
  – (1) Confirmed, or  
  – (2) Excluded (missing or invalid criteria).

E) SUBAGENT TASK SUMMARY  
Below is the step-by-step workflow using the available skills. The numbering here reflects the logical sequence of checks for each business entry:

1) For each business record in the subagent input, do the following:

(a) Domain Launch Date & Authority Check  
       • Use #224 - Oracle Ask A Question to look up the domain’s launch date (search official government databases, business listings, domain records), check domain authority (e.g., SEMRush, other SEO tools if feasible).  
         INPUT: A direct question like “Find the domain creation/launch date and domain authority for [Website URL].”  
         OUTPUT: The Oracle’s response with domain age data, authority score, any other relevant domain analysis.

(b) Validate Business Launch Criteria  
       • Evaluate if the business was launched within the past 3 years according to the Oracle’s returned data.  
       • Evaluate if domain authority < 30.  
       • If either fails, mark as “excluded.”

(c) Verify Website Functionality  
       • Use #224 - Oracle Ask A Question again, if needed, to confirm that the website is active/functional and not returning HTTP errors. E.g., “Check if [Website URL] is currently live and functional.”  
       • If the website is offline or broken, “excluded.”

(d) Check Traffic / Web Visibility  
       • (Optional) In the same Oracle question or a subsequent one, “Estimate if [Website URL] has minimal or no traffic and is hard to find in Google/Bing.”  
       • If the site is robustly visible, skip or “excluded.”

(e) Validate Contact Person & Contact Info  
       • Use #224 - Oracle Ask A Question or #226 - Extract Structured Data From 1x URL (if needed) to confirm that the named contact person is indeed associated with the business.  
       • Check the provided email address for a valid format or known existence (the Oracle can do a quick ping or hush check: “Check if [Email Address] is valid and associated with [Website/Company Name].”).  
       • Check phone number format or note if it appears disconnected.  
       • If any info is missing or found invalid, mark the field as “blank” or “excluded.”

(f) Assess Social or Additional Contact Links  
       • If additional social links are present, use #224 - Oracle Ask A Question (if desired) to confirm if they belong to the same business or contact person.  
       • If unknown, leave blank but do not exclude the record if all other validations pass.

(g) Final Decision  
       • If all the above checks pass, store the record as “included” with updated/verified details.  
       • If any fails, record is “excluded.”

2) After iterating through all records, compile the validated results.

3) Return the updated dataset (with included/excluded statuses) as the subagent’s final output.

F) SILOS  
Below is a simplified silo breakdown if needed:

• Silo 1: Domain & Launch Check  
   – Step (a) Domain creation date / authority  
   – Step (b) Launch date <3 yrs?

• Silo 2: Website & Traffic Check  
   – Step (c) Website functionality  
   – Step (d) Traffic / minimal presence

• Silo 3: Contact Validation  
   – Step (e) Confirm contact person & phone/email  
   – Step (f) Check additional links

• Silo 4: Final Acceptance or Exclusion  
   – Step (g) Include or exclude the record

This completes the subagent’s workflow for re-checking and filtering the domain, business criteria, and contact details.

5 Template & Links

Expand Flow

A) SUBAGENT SUMMARY  
This subagent focuses on verifying the viability of each recently discovered business (from a prior step), ensuring that each meets the strict criteria (recently launched, low or no domain authority, minimal visibility, valid contact info, etc.) before passing along final validated data.

B) FINAL TASK OUTPUT  
A filtered, validated dataset (e.g., JSON or similar structured data) containing only those businesses that meet all “recently launched and low domain authority” requirements, along with confirmed contact information.

C) SUBAGENT INPUT  
• A structured list of candidate businesses (URL, domain info if available, contact name, email, phone number, potential social links, or any other data discovered in prior steps).  
• Criteria for validation (e.g., domain authority < 30, launched ≤ 3 years ago, minimal social presence, etc.).

D) SUBAGENT TASK SUMMARY  
Below is the complete flow of how this subagent might operate, describing how each skill is chained:

1) Subagent Receives Candidate Data  
   Input: “structured list of candidate businesses plus criteria”

2) For Each Business, Verify Domain Authority & Domain Launch Date  
   • Skill #224 – “Oracle Ask A Question”  
     – Action: Query external sources (e.g., SEMrush, government business databases, social posts, etc.) to check:  
       (a) Domain authority score (is it < 30, minimal backlinks, etc.?)  
       (b) Confirm approximate age/launched date (is it ≤ 3 years old?)  
       (c) Confirm minimal social presence (small or negligible follower counts).  
     – Output: Answers to domain authority, launch date, and social presence.

3) Re-Check Contact Validity  
   • Skill #223 – “Powerful LLM Prompt-to-Text Response”  
     – Action: For each business’s contact data, quickly confirm that the main contact email and phone number appear genuine, or use additional queries (if needed) to see if the email/phone is obviously invalid.  
     – Output: A short validity label or commentary for each contact piece (e.g., “verified / uncertain / invalid”).

4) Filter Out Non-Compliant Entries  
   • No direct skill call here—this is purely an internal logic step:  
     – If domain authority ≥ 30 → discard.  
     – If business age > 3 years → discard.  
     – If contact data obviously invalid → discard.

5) Return Validated List  
   Subagent Final Output: A refined, validated listing of businesses that pass all checks, with updated confidence markers for domain authority, age, contact details, and social presence.

E) SILOS  
• SILO 1 – Domain & Age Verification  
   – Use skill #224 (Oracle) to gather domain authority, business launch date, and minimal social presence.  
• SILO 2 – Contact Validation  
   – Use skill #223 (Powerful LLM) or additional checks with #224 to confirm email/phone legitimacy.  
• SILO 3 – Final Filter & Return  
   – Internally remove businesses that fail key thresholds, then compile final validated list with all details.