What Shall We Build Next?

Describe

Describe your task

Refine

Refine the plan

SubAgents

Review all agents

Deploy

Deploy your agent

A) SUBAGENT SUMMARY  
Retrieve token data from gmgn.ai, then apply filtering rules (liquidity ≥100,000; volume ≥200,000; token age ≤48 hours; holders ≥500) to produce a final “filtered-token-list.”

──────────────────────────────────
B) FINAL TASK OUTPUT  
A final filtered list of tokens (in JSON or text form) that satisfy the required minimum thresholds for liquidity, volume, age, and holder count.

──────────────────────────────────
C) SUBAGENT INPUT  
• The gmgn.ai URL (points to a page or specific endpoint with token data).  
• Instructions on which token fields to extract (liquidity, volume, token age, holder count).  
• The numeric filter thresholds:
  – Liquidity ≥100,000  
  – Volume ≥200,000  
  – Token age ≤48 hours  
  – Holders ≥500

──────────────────────────────────
D) SUBAGENT TASK SUMMARY  
1) Skill #226 – Extract Structured Data From 1x URL  
   – Input: (URL to gmgn.ai, plus what data to extract: liquidity, volume, token age, holder count)  
   – Output: A structured summary of all token data found on gmgn.ai (returned in text form, typically key-value style).

2) Skill #223 – Powerful LLM Prompt-to-Text Response  
   – Input: The structured token data from Skill #226, plus a prompt that explains the filter criteria (liquidity ≥100,000; volume ≥200,000; token age ≤48 hours; holders ≥500).  
   – Output: A filtered list (JSON or text) of tokens that meet all criteria.

3) Subagent Final Output  
   – The filtered token list [“filtered-token-list”].

──────────────────────────────────
E) SILOS  
SILO 1: Fetch Data  
• Action: Skill #226  
• Goal: Gather raw token data about liquidity, volume, token age, and holder count from gmgn.ai.

SILO 2: Filter Data  
• Action: Skill #223  
• Goal: Parse and apply the filtering rules (≥100k liquidity, ≥200k volume, ≤48 hours age, ≥500 holders) to produce the final “filtered-token-list.”

──────────────────────────────────
END OF SUBAGENT 1 FLOW

SubAgent #1 - Diagram

Expand Diagram

A) SUBAGENT SUMMARY  
The “DueDiligenceAgent” checks each filtered token’s contract quality, holder distribution, and social media presence, then outputs a pass/fail or risk label per token.

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

B) FINAL TASK OUTPUT  
A single structured report (e.g., JSON) that appends the following info to each token:  
• Contract Quality (label from RugcheckXYZ)  
• Holder Distribution Summary (BubbleMaps)  
• Social Media Mentions (TweetScout)  
• A Consolidated Pass/Fail/Risk Rating

Example format (JSON/text table):  
[
  {
    "token": "TokenName",
    "contractAddress": "0xABC123...",
    "rugcheckRating": "Good",
    "holderDistribution": "...",
    "socialMediaPresence": "Score or summary from TweetScout",
    "finalVerdict": "Pass"   // or "Fail" or "Risk"
  },
  ...
]

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

C) SUBAGENT INPUT  
A list of filtered tokens from the previous subagent, where each token entry has at least:  
• token name or ticker,  
• contract address,  
• basic overview (optional).

Example:  
[
  {
    "token": "ExampleToken",
    "contractAddress": "0x1234abc...",
    "otherDetails": ...
  },
  ...
]

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

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

1) Receive the [filtered-token-list] as input.

2) FOR EACH token in the [filtered-token-list], perform these three Oracle queries (repeating Skill #224 up to 3 times):

(a) Skill #224 – Oracle Ask A Question  
       • PURPOSE: Ask “Is this token labeled ‘Good’ or flagged in any way by @Rugcheckxyz? Provide results for contract address X.”  
       • INPUT: The token’s contract address, plus the question about contract quality.  
       • OUTPUT: Response about the token’s contract rating or any known flags.

(b) Skill #224 – Oracle Ask A Question  
       • PURPOSE: Ask “What does BubbleMaps say about holder distribution for contract address X? Summarize any concentration or red flags.”  
       • INPUT: The same token’s contract address, plus the question about holder distribution.  
       • OUTPUT: Response about unusual concentration or distribution patterns.

(c) Skill #224 – Oracle Ask A Question  
       • PURPOSE: Ask “Does @TweetScout_io or social media mention or discuss this token? Summarize sentiment, popularity, or any red flags for contract address X.”  
       • INPUT: The same token’s contract address, plus the question about social presence.  
       • OUTPUT: Response about social media coverage or sentiment.

3) Consolidate the 3 responses for each token:

(a) Skill #223 – Powerful LLM Prompt-to-Text Response  
       • PURPOSE: Merge the three Oracle responses and generate a final label or rating (Pass, Fail, or Risk). For example, if contract is “Good,” holder distribution is fair, and social presence is not concerning, label “Pass.” If major issues appear, label as “Fail” or “Risk.”  
       • INPUT: The three responses (from step 2), plus logic for deciding pass/fail/risk.  
       • OUTPUT: A consolidated block of text or structured snippet with final verdict appended to the token’s record.

4) Accumulate the enriched records (with contract rating, distribution summary, social media mention, final pass/fail/risk) into a single data structure—this is the [due-diligence-report].

5) Return the [due-diligence-report] as the final subagent output.

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

F) SILOS

SILO 1: ORACLE QUERIES  
• For each token, call Skill #224 up to three times with different question contexts (RugCheck, BubbleMaps, TweetScout).

SILO 2: CONSOLIDATION & VERDICT  
• Use Skill #223 to merge the three Oracle responses for each token and produce the final pass/fail/risk rating.  
• Gather all tokens into a single data structure.

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

This completes the subagent flow for the DueDiligenceAgent.

SubAgent #2 - Diagram

Expand Diagram

A) SUBAGENT SUMMARY  
Generates a Python code snippet that can send a token’s contract address and buy parameters to the Toxi Solana Bot, returning logs on success or failure.

B) FINAL TASK OUTPUT  
A plain-text Python code snippet (string output) that interacts with Toxi Solana Bot given the contract address, buy amount, and relevant method of integration.

C) SUBAGENT INPUT  
• Toxi Solana Bot usage instructions (if available)  
• Token contract address (string)  
• Purchase amount (float or integer)  
• Any additional parameters required by the sniper bot (e.g., slippage, user authentication)

E) SUBAGENT TASK SUMMARY  
1) The subagent receives as input:  
   – The Toxi Solana Bot usage instructions (if applicable)  
   – The token’s contract address to purchase  
   – The desired purchase amount

2) It then uses Skill #223 – Powerful LLM Prompt-to-Text Response to generate a Python code snippet. This snippet should contain:  
   • The required function (or class) to interact with Toxi Solana Bot  
   • Parameters for contract address, buy amount, and optional slippage  
   • Logging or print statements to confirm success or failure

3) The subagent returns the generated code.

Flow Diagram:  
   • [subagent input] → (Skill #223 with prompt: “Generate the Toxi Solana Bot integration code…”) → [Python code snippet]

F) SILOS  
Given the simplicity of the requirement, the subagent operates in a single silo:

SILO 1 – Generate Python Sniper Integration Code  
• Input: Toxi Bot instructions, contract address, and buy amount  
• Process: Use Skill #223 to produce the Python code snippet  
• Output: Python code (text string)

SubAgent #3 - Diagram

Expand Flow

A) SUBAGENT SUMMARY  
────────────────────────────────────────────────────────
This subagent fetches newly listed memecoins from PumpFun and gmgn.ai, checks their contract status (e.g., “Good” vs. “Bad”), applies a minimal risk/quality score, and returns a final filtered list of promising memecoins.

B) FINAL TASK OUTPUT  
────────────────────────────────────────────────────────
A single final dataset (e.g., JSON or text) containing each discovered memecoin’s relevant data (contract address, volume, age, status label) along with a pass/fail rating or “score.” This is the “[memecoin-discovery-list].”

C) SUBAGENT INPUT  
────────────────────────────────────────────────────────
• URLs for PumpFun and gmgn.ai endpoints (where new memecoins are listed).  
• Any filtering instructions or thresholds to help decide pass/fail (e.g., “Ignore tokens older than 48 hours” or “Only include tokens labeled as Good by Rugcheckxyz”).

D) SUBAGENT TASK SUMMARY  
────────────────────────────────────────────────────────
Below is the step-by-step workflow showing the Subagent’s actions and how the outputs from each skill feed into the next step.

1. Skill #226 – Extract Structured Data From 1x URL  
   • INPUT: PumpFun URL and instructions to “Fetch newly listed memecoins, returning contract address, volume, age, etc.”  
   • OUTPUT: JSON/text with newly listed memecoins’ data from PumpFun.

2. Skill #226 – Extract Structured Data From 1x URL (again)  
   • INPUT: gmgn.ai URL and instructions to “Fetch newly listed memecoins, returning contract address, volume, age, etc.”  
   • OUTPUT: JSON/text with newly listed memecoins’ data from gmgn.ai.

3. Skill #223 – Powerful LLM Prompt-to-Text Response  
   • PURPOSE: Merge the data from both PumpFun and gmgn.ai so we have a combined “memecoin candidate” dataset.  
   • INPUT: The two data outputs from Skills #226 (PumpFun + gmgn.ai).  
   • OUTPUT: A single merged list of memecoins, removing duplicates.

4. Skill #224 – Oracle Ask A Question (loop or batch call)  
   • PURPOSE: For each candidate memecoin, query external resources (e.g., “@Rugcheckxyz, is [contract_address] flagged Good or Bad?”).  
   • INPUT: Merged memecoin data from step 3, contract addresses or tickers, plus question about contract status.  
   • OUTPUT: A text or JSON object summarizing each memecoin’s contract risk status.

5. Skill #223 – Powerful LLM Prompt-to-Text Response  
   • PURPOSE: Evaluate each token based on region data, historical analysis (if available), “Good” vs. “Bad” contract label, volume/holder criteria, or other minimal risk checks. Assign a pass/fail or a success score.  
   • INPUT: Merged candidate list from step 3 + the contract status info from step 4 + user’s pass/fail rules.  
   • OUTPUT: Final list of promising memecoins with pass/fail or score appended. This is the subagent’s final output: “[memecoin-discovery-list].”

E) SILOS (IF NEEDED)  
────────────────────────────────────────────────────────
• Silo 1: Data Gathering  
  - Steps 1 & 2 (Skill #226) to fetch memecoin data from PumpFun and gmgn.ai.  
• Silo 2: Data Merging  
  - Step 3 (Skill #223) to unify the lists into a single set of candidate memecoins.  
• Silo 3: Contract Checks  
  - Step 4 (Skill #224) to call external APIs (e.g., Rugcheckxyz) for “Good” vs. “Bad.”  
• Silo 4: Final Filtering/Scoring  
  - Step 5 (Skill #223) to apply logic that yields a pass/fail or risk score.

The output from Step 5 is the subagent’s final deliverable: a consolidated “memecoin-discovery-list” of newly discovered, promising memecoins.

4 Template & Links

Expand Flow

A) SUBAGENT SUMMARY  
This subagent (WorkflowOrchestrator) chains together the outputs of the other subagents (DataFetcherFilter → DueDiligenceAgent → SniperIntegrationAgent → MemecoinDiscoveryAgent) into one daily automated process. It generates the single Python “main” script that runs these steps in sequence, providing logs and status updates along the way.

B) FINAL TASK OUTPUT  
• A consolidated Python script (plain text) containing one “main” function (or similar) that, when run, executes the full daily workflow in order:  
  1) Fetch and filter tokens from gmgn.ai.  
  2) Perform due diligence checks (contract quality, holders, social presence).  
  3) If passing, invoke sniper bot purchase.  
  4) Scan for new memecoins, do minimal checks, and integrate potential buys.  
• This final script is the [full-workflow-script] and is output as text.

C) SUBAGENT INPUT  
This subagent needs:  
• User configuration and scheduling details (e.g., daily run time, risk management parameters).  
• Any relevant credentials or API keys (wallet, Toxi Solana Bot, etc.).  
• Filter parameters (liquidity, volume, etc.).  
• Risk thresholds and allowable user trade settings.

E) SUBAGENT TASK SUMMARY  
Below is the step-by-step logic the subagent follows to build the final script. Where necessary, it calls on Skill #223 (“Powerful LLM Prompt-to-Text Response”) to generate or consolidate Python code that orchestrates all tasks:

1) Subagent Setup (Generate Orchestration Logic)  
   • (Skill #223) Generate cohesive Python scaffolding that will include:  
     – Imports, environment variables, scheduling.  
     – A “main” function controlling the daily run.  
     – Pseudocode placeholders representing each subagent’s function.

2) Integrate Subagent 1 (DataFetcherFilter)  
   • (Skill #223) Insert or call the snippet that:  
     – Uses Skill #226 internally to fetch data from gmgn.ai.  
     – Uses Skill #223 to filter tokens by liquidity, volume, age, holders.  
     – Returns [filtered-token-list].  
   • Log or store the filtered list for later steps.

3) Integrate Subagent 2 (DueDiligenceAgent)  
   • (Skill #223) Insert or call the snippet that:  
     – For each token in [filtered-token-list], uses Skill #224 to confirm contract quality via Rugcheckxyz, holder distribution on bubblemaps, social presence on TweetScout_io.  
     – Uses Skill #223 to finalize pass/fail/risk labeling.  
     – Returns [due-diligence-report].  
   • Log or store which tokens passed.

4) Integrate Subagent 3 (SniperIntegrationAgent)  
   • (Skill #223) Insert or call the snippet that:  
     – For each “passed” token, sends the contract address (CA) to Toxi Solana Bot.  
     – Triggers buy commands and logs the result.  
     – Optionally handle any purchase logic (limited to 20% capital usage, or user-defined).

5) Integrate Subagent 4 (MemecoinDiscoveryAgent)  
   • (Skill #223) Insert or call the snippet that:  
     – Uses Skill #226 and Skill #224 to fetch newly listed memecoins and check contracts.  
     – Uses Skill #223 to finalize a success score or pass/fail.  
     – Optionally triggers the same buy logic for promising memecoins.  
   • Log or store [memecoin-discovery-list] results.

6) Logging & Final Assembly  
   • (Skill #223) Combine all steps into one consistent Python file.  
     – Provide summary logs of tokens bought, tokens analyzed, memecoins discovered.  
     – Include all error handling, schedule logic, user parameters (exit strategies, risk management, running frequency).  
     – Provide hooks to subagent #6’s instructions or comments for environment setup.

7) Output the Final Script  
   • (Skill #223) Return the complete Python script text as [full-workflow-script].

F) SILOS

SILO 1: Create Main Workflow Structure  
• Inputs: user config (daily schedule, etc.).  
• Action: (Skill #223) → generate “main” function skeleton with Python code.

SILO 2: Call DataFetcherFilter Subagent  
• Action: (Skill #223) → embed logic to fetch & filter.  
• Output: [filtered-token-list].

SILO 3: Call DueDiligenceAgent Subagent  
• Action: (Skill #223) → embed logic for contract checks, holder analysis, social presence checks.  
• Output: [due-diligence-report].

SILO 4: Call SniperIntegrationAgent  
• Action: (Skill #223) → embed logic to buy tokens that pass due diligence.  
• Output: purchase confirmations/logs.

SILO 5: Call MemecoinDiscoveryAgent  
• Action: (Skill #223) → embed logic to check new gems, incorporate buy logic if greenlit.  
• Output: [memecoin-discovery-list].

SILO 6: Final Assembly & Return  
• Action: (Skill #223) → merges all code into a single Python script, logs final results, returns the entire text as the subagent’s final output.

5 Template & Links

Expand Flow

A) SUBAGENT SUMMARY  
Generates a concise “Setup Instructions” text (e.g. environment setup, library requirements, credential placement, scheduling instructions) for the trading bot.

B) FINAL TASK OUTPUT  
A single block of text (setup-instructions) explaining all necessary environment prerequisites, library installations, configuration steps, and scheduling options.

C) SUBAGENT INPUT  
• (Optional) Basic user environment details (operating system, Python version, scheduling preferences) to tailor the instructions, OR none if the agent should produce generic setup instructions.

E) SUBAGENT TASK SUMMARY  
1. (Subagent Input) → 2. (Use Skill #223 – “Powerful LLM Prompt-to-Text Response”) → 3. (Subagent Output: setup-instructions)

Detailed Flow:  
1) Receive (optional) environment details and instructions about what to include in setup guidelines.  
2) Use Skill #223 to generate a detailed text tutorial covering environment setup and libraries, where to store wallet credentials, how to schedule daily runs (e.g. cron on Linux/Mac, Task Scheduler on Windows), any additional user instructions for configuring the bot, and steps for verifying everything is properly installed.  
3) Collect the resulting text and return it as the final [setup-instructions].

F) SILOS  
• Silo 1: Environment Setup Instructions Generation  
  – Input: (Optional) environment details or blank.  
  – Skill: #223 (LLM prompt) to produce the comprehensive setup text.  
  – Output: final text instructions (setup-instructions).