The Asynchronus Agents

The intelligence and capability of the Asynchronus ecosystem stem from its specialized AI sub-agents. These are the "workers" who execute the tasks you delegate via Shell. Each agent is an expert in its domain, equipped with specific tools and knowledge. They are built to operate autonomously yet collaboratively, orchestrated by Shell and powered by the Graph framework.

Think of them as members of an elite, decentralized Web3 operations team, ready to tackle any challenge.

Introduction to Sub-Agents

Before we meet the individual agents, let's understand their shared characteristics:

• AI-Powered: Each agent leverages AI models for analysis, decision-making, and execution within its specialty.

• Domain-Specific: They possess deep knowledge and tailored tools for their area (e.g., cross-chain bridging, security analysis).

• Asynchronous: They can work independently and in parallel, drastically speeding up complex workflows that involve multiple steps or waiting periods.

• Communicative: They interact through shared memory and messaging, allowing them to pass data, coordinate actions, and work towards a common goal defined by Shell.

• Composable: They are designed to work together, as we'll see in the "Composable Agents" section.

Let's meet the current members of the team:

✅ Analyze Agent

• Description: The research and intelligence arm of Asynchronus. This AI assistant excels at gathering, processing, and interpreting vast amounts of data from across protocols, markets, and chains to provide actionable insights.

• Capabilities:

  • Market Research: Identifying trends, tracking narratives, and understanding broad market dynamics.

  • Token Analysis: Deep dives into specific tokens, including fundamentals, on-chain metrics, tokenomics, and holder distribution.

  • Ecosystem Trends: Evaluating the health, growth, and activity within specific L1/L2 ecosystems or DeFi sectors.

  • Sentiment Analysis: Gauging community and market feeling towards projects, tokens, or events.

  • Wallet/Portfolio Analysis: Examining the contents, history, and performance of specific Web3 wallets.

• Use Cases:

  • "Find me undervalued AI-related tokens on the Arbitrum network."

  • "What is the general sentiment towards the upcoming EIP-4844 upgrade?"

  • "Compare the daily active users and transaction volumes of GMX and Synthetix over the last 30 days."

  • "Give me a risk/reward profile for the PENDLE token."

✅ Bridge Agent

• Description: The cross-chain logistics expert. This agent is designed to facilitate secure and efficient transactions and interoperability across the increasingly fragmented landscape of L1 and L2 blockchains.

• Capabilities:

  • Asset Bridging: Executing the transfer of tokens from one blockchain network to another.

  • Protocol Routing: Identifying the most efficient (cost, speed, security) bridging protocols or routes for a given transfer.

  • Transaction Cost Comparison: Analyzing gas fees and protocol fees across different bridging options.

• Use Cases:

  • "What's the cheapest way to send 1 WETH from Ethereum to Base?"

  • "Bridge my 500 USDC on Polygon to Optimism."

  • "Help me move liquidity from Solana to Arbitrum."

  • "Estimate the time it will take to bridge USDT via Stargate vs. Router Protocol."

✅ Swap Agent

• Description: The on-chain trading specialist. This AI assistant focuses on executing token swaps within the same blockchain, optimizing for the best possible execution on Decentralized Exchanges (DEXs).

• Capabilities:

  • Token Swapping: Performing A-to-B token trades on various DEXs.

  • Slippage Optimization: Finding routes and execution strategies to minimize price impact, especially for larger trades.

  • Trade Execution: Interacting directly with DEX smart contracts to perform swaps.

  • Best Rate Finding: Querying multiple DEX aggregators and individual DEXs to find the optimal price.

• Use Cases:

  • "Swap 1 ETH for USDC on Arbitrum, ensuring I get the best rate."

  • "Sell 10,000 ARB for DAI, but ensure slippage doesn't exceed 0.5%."

  • "Is it cheaper to swap CRV to ETH via Uniswap or Sushiswap right now?"

  • "Execute a limit order to buy WBTC when its price drops below $65,000 (USDC)." (Note: Advanced features like limit orders may depend on specific protocol integrations)

✅ Prediction Agent

• Description: The forward-looking analyst. This sub-agent leverages statistical models, machine learning, and historical data to provide predictive insights and forecasts.

• Capabilities:

  • Price Trend Predictions: Forecasting potential short-term and long-term price movements for various assets.

  • Volatility Analysis: Assessing and predicting periods of high or low market volatility.

  • Model-Based Forecasting: Utilizing specific, trained models to generate quantitative predictions (e.g., APY forecasts, risk scores).

• Use Cases:

  • "What's the predicted price range for SOL over the next 7 days?"

  • "Is the market likely to experience high volatility next week?"

  • "Based on current trends, forecast the potential yield for the Lido stETH pool."

  • "Which tokens show a high probability of upward movement based on current indicators?"

✅ Security Agent

• Description: The vigilant guardian of your Web3 activities. This sub-agent is dedicated to identifying risks, assessing smart contracts, and ensuring the safety of your interactions.

• Capabilities:

  • Contract Risk Profiling: Analyzing smart contracts for known vulnerabilities, suspicious code, and centralization risks.

  • Honeypot Detection: Identifying potential scam tokens or protocols designed to trap user funds.

  • Contract Audits: Checking if a protocol has undergone formal security audits and summarizing their findings.

  • Protocol Safety Metrics: Evaluating overall protocol health based on factors like TVL, age, audit history, and on-chain activity.

• Use Cases:

  • "Is this new DeFi protocol (contract address 0x...) safe to interact with?"

  • "Check if this token (contract address 0x...) looks like a honeypot."

  • "What are the main security risks associated with EigenLayer restaking?"

  • "Before I sign this transaction, can you do a quick security check on the destination contract?"

Composable Agents: How They Work Together

The true power of Asynchronus emerges when these specialized agents compose their abilities to execute complex, multi-step workflows. Shell acts as the conductor, but the agents are the orchestra, playing their parts in harmony.

Consider this workflow: "Find the highest-yield, secure lending pool for my 10 ETH on L2s, and move my funds there."

Here's how the agents might collaborate:

1. Shell: Receives and understands the user's intent.

2. Analyze Agent: Scans known L2s (Arbitrum, Optimism, Base, etc.) and their lending protocols (Aave, Compound, etc.) to find the current highest APYs for ETH. It lists the top 3 options.

3. Prediction Agent: Looks at the historical APY data for the top 3 and forecasts their likely stability over the next month.

4. Security Agent: Vets the smart contracts and overall protocol health of the top 3 options. It flags one as being too new and lacking audits.

5. Shell: Presents the remaining two, highly-rated options to the user, along with the analysis.

6. User: Selects Option A on Arbitrum.

7. Shell: Confirms the user's ETH is currently on Ethereum Mainnet.

8. Bridge Agent: Finds the most efficient and cost-effective bridge to move 10 ETH from Mainnet to Arbitrum.

9. Shell: Presents the bridging transaction to the user for approval.

10. User: Approves.

11. Bridge Agent: Executes the bridge. Shell monitors for completion.

12. Shell: Confirms ETH arrival on Arbitrum.

13. Swap/Action Agent: Prepares the transaction to deposit the 10 ETH into the selected Aave pool on Arbitrum.

14. Shell: Presents the lending transaction to the user for approval.

15. User: Approves.

16. Swap/Action Agent: Executes the deposit.

17. Shell: Confirms the successful deposit and reports back to the user: "Your 10 ETH is now successfully deposited and earning yield in Aave on Arbitrum."

This seamless orchestration—querying data, analyzing risk, moving assets across chains, and interacting with protocols—is made possible by the composable nature of Asynchronus agents, all managed through a simple conversational prompt. As more agents are developed and added via the Graph framework, the complexity and power of these workflows will only grow.