AuditSwap: Autonomous Risk-Gated Liquidity Layer

An autonomous, end-to-end financial engine that leverages Birdeye's real-time on-chain data to discover newly deployed Solana tokens, mathematically verify their security profiles, and immediately execute programmatic market placements.

By removing human latency and emotional bias, AuditSwap converts raw blockchain signals into secure, structured capital deployment within milliseconds.

---

⚡ The Core Problem

On high-throughput networks like Solana, thousands of new tokens are listed daily. This presents a massive dilemma for capital allocators:

• The Latency Trap: Acting fast enough to capture early market momentum usually means buying blindly without verifying contract safety.
• The Security Trap: Taking the time to manually audit a contract for malicious code or dangerous holder concentration means missing the optimal entry window.

Human latency gets traders rugged; traditional bots get trapped.

---

🛠️ The Birdeye Solution

AuditSwap eliminates the trade-off between speed and security by building a direct pipeline between real-time data and atomic settlement.

The engine continuously processes the live chain state, evaluating risks programmatically using Birdeye's enterprise infrastructure.

Diagram 1 — End-to-End Pipeline

flowchart LR
    A[Solana Blockchain] -->|New token mints| B[Birdeye /new_listing]
    B -->|Unfiltered token addresses| C[Nexus Engine]
    C -->|Request security metrics| D[Birdeye /token_security]
    D -->|Risk scores & flags| C
    C -->|Audit decision: PASS| E[Jupiter v6 API]
    E -->|Quote & route| F[Atomic Swap Execution]
    F -->|Swap complete| G[User / Treasury]
    
    C -->|Audit decision: FAIL| H[Reject & Log]
    H -->|No swap| G

Loading

---

Diagram 2 — Security Decision Tree (Audit Filter)

This is the core audit filter logic — sequential gates, fail any = abort.

flowchart TD
    START([New token detected via /new_listing]) --> A{Mint authority present?}
    
    A -->|Yes| REJECT1[ REJECT: Centralized mint risk]
    REJECT1 --> ABORT([Abort swap])
    
    A -->|No| B{Honeypot simulation score < threshold?}
    
    B -->|No| REJECT2[ REJECT: Honeypot risk]
    REJECT2 --> ABORT
    
    B -->|Yes| C{Top 10 holders concentration < 30%?}
    
    C -->|No| REJECT3[ REJECT: Whale concentration risk]
    REJECT3 --> ABORT
    
    C -->|Yes| D{Liquidity > $50k?}
    
    D -->|No| REJECT4[ REJECT: Insufficient liquidity]
    REJECT4 --> ABORT
    
    D -->|Yes| PASS([ PASS: Eligible for swap])
    PASS --> EXECUTE[Execute via Jupiter v6]

Loading

---

Diagram 3 — Birdeye Metric Mapping to Nexus Engine

Shows exactly which Birdeye endpoints are consumed and how they map into the audit decision.

flowchart LR
    subgraph BIRDEYE["Birdeye API"]
        NL["/v2/tokens/new_listing\n(real-time new token stream)"]
        TS["/defi/token_security\n(risk scores, mint authority, honeypot, holders)"]
        TP["/defi/token_overview\n(price, liquidity, volume)"]
    end

    subgraph NEXUS["AuditSwap Engine"]
        A1["Parse token address"]
        A2["Aggregate security flags"]
        A3["Run decision tree"]
        A4["Output: PASS / FAIL"]
    end

    subgraph JUP["Jupiter v6"]
        J1["/quote"]
        J2["/swap"]
    end

    NL -->|Token address| A1
    TS -->|Mint authority, honeypot, holders| A2
    TP -->|Liquidity, market cap| A2
    A2 --> A3
    A3 -->|If PASS| J1
    J1 --> J2
    J2 -->|Atomic swap| EXECUTION[On-chain swap]

Loading

---

Evaluation Architecture

The agent treats the incoming market stream as a sequential state system, analyzing risk metrics objectively to guarantee programmable capital guardrails:

Metric Evaluated	Birdeye Endpoint Source	Programmatic Action
Token Genesis	/v2/tokens/new_listing	Ingests contract address for immediate analysis.
Honeypot Risk	/defi/token_security	Immediate abort if malicious transfer functions exist.
Mint Authority	/defi/token_security	Flags and drops if mint is not renounced.
Ownership Concentration	/defi/token_security	Flags and drops if top 10 holders hold >30% supply.
Liquidity Verification	/defi/token_security & /defi/token_overview	Ensures pool liquidity > $50k.

Audit Filter Summary Table

Gate	Condition	Data Source	Action on Fail
1	Mint authority == null	Birdeye /defi/token_security	Reject
2	Honeypot score < 0.05	Birdeye simulation	Reject
3	Top 10 holders ≤ 30%	Birdeye holder distribution	Reject
4	Liquidity > $50k	Birdeye /defi/token_overview	Reject
All gates pass	→	→	Execute Jupiter swap

---

Quick Start & Deployment

Prerequisites

• Node.js v18+ & TypeScript
• A Birdeye Enterprise or Developer API Key
• A Solana RPC URL and a funded keypair

1. Installation

git clone https://github.com/yourusername/nexus-router.git
cd nexus-router
npm install

2. Environment Configuration

Create a .env file in the root directory:

BIRDEYE_API_KEY=your_birdeye_api_key_here
RPC_URL=https://api.mainnet-beta.solana.com
PRIVATE_KEY=your_base58_encoded_secret_key
JUPITER_API=https://quote-api.jup.ag/v6

3. Run the Autonomous Loop

npx ts-node agent.ts

Run Tests

npm test

---

Repository Structure

nexus-router/
├── src/
│   ├── birdeye/      # Birdeye API clients (/new_listing, /token_security)
│   ├── audit/        # Decision tree logic (4-gate filter)
│   ├── jupiter/      # Jupiter v6 integration
│   └── listener/     # Real-time /new_listing consumer
├── tests/            # Unit + integration tests
├── config/           # Threshold configuration (liquidity, holder %, etc.)
├── agent.ts          # Main autonomous loop entrypoint
└── README.md

---

Hackathon Qualification Checklist

• Product Utility: Protects real capital by filtering out malicious token launches programmatically.
• Technical Depth: Combines read-only market analytics with low-latency write actions via decentralized liquidity routers.
• Birdeye Integration: Actively integrates both /v2/tokens/new_listing and /defi/token_security data.
• High-Volume Pinging: Continuously loops to comfortably surpass the 50+ API calls threshold required for submission.

---

Contributing

See CONTRIBUTING.md for audit filter contribution guidelines.

---

License

MIT — see LICENSE for details.

How to use this for your final submission today

1. Save this exact text as README.md in your GitHub repository.
2. Ensure your agent.ts file has run long enough to clear the 50 API call quota on Birdeye's developer dashboard.
3. Record a clean, quick terminal screenshot video of the script running, then post your submission tweet using the #BirdeyeAPI hashtag.

---

**What's included:**

- ✅ All three Mermaid diagrams (live-rendering on GitHub)
- ✅ Your original "Core Problem" + "Birdeye Solution" sections
- ✅ The Evaluation Architecture table
- ✅ The Audit Filter Summary Table
- ✅ Quick Start + Environment setup
- ✅ Repository structure
- ✅ Hackathon checklist
- ✅ Submission instructions

You can now **copy the raw markdown above** directly into your `README.md`. The diagrams will render automatically when viewed on GitHub.