-
-
Save luckysitara/90b3c5587a6137831658a92ad9826166 to your computer and use it in GitHub Desktop.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| location wasn't included in the ip analysis output( and also if the analysis return the latitude and longitude of the location is there a way we can add a map feature using google map to view the location? ) and the result instead of being displayed in the same page can have a different page for it just like pii intelligence result page | |
| https://v0.app/chat/sentinel-6pqhUTAgEUc | |
| AIzaSyCoNJkO0GhoeY2uy6NPzxeVF0GxZvy33qI |
Author
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
This is an excellent next step. A well-defined Product Requirements Document (PRD) is the blueprint for a winning hackathon submission, ensuring your team is aligned on scope, features, and success metrics.
Here is a detailed PRD for your project, 'Solana Copilot'.
📝 Product Requirements Document (PRD)
Project: Solana Copilot (AI-Powered DeFi Agent)
🎯 Goal
To build a secure, demonstrably working Minimum Viable Product (MVP) of an AI Agent on Solana that simplifies complex, multi-step DeFi transactions using natural language, drastically reducing friction for users.
❓ Problem Statement
Executing multi-step financial actions on Solana DeFi (e.g., swapping tokens, providing liquidity, staking LSTs) requires users to navigate multiple dApps, understand complex Web3 jargon (slippage, impermanent loss, epochs), and manually calculate required token inputs. This complexity creates a steep barrier to entry, leading to confusion, errors, and lost capital.
✅ Success Metrics (KPIs for Judging)
| Metric | Target | Notes |
|---|---|---|
| P0: Transaction Chain Success Rate | 100% | The Agent must successfully execute a minimum of one 3-step transaction chain (e.g., Swap, LP Deposit, Stake). |
| P1: Interpretive Accuracy | 90% | The Agent's pre-execution summary must accurately describe the on-chain action in plain English for 9/10 unique prompts. |
| P2: Core Tool Integration | Min. 3 | Successfully integrate and leverage three key Solana DeFi protocols (Jupiter, Orca, Kamino/Marinade). |
| P3: Wallet Insight Utilization | 100% | Agent must always reference the user's current token balances in its reasoning before proposing an action. |
👤 Target User: The DeFi Novice/Efficiency Seeker
🚀 User Stories (The Core MVP Scope)
| Priority | User Story | Agent Capability Focus |
|---|---|---|
| P0 (Critical) | As a connected user, I want to type "Deposit all my USDC into the highest-yielding Orca pool," so that the Agent autonomously finds the pool and generates the transaction. | Multi-Step Tool Use |
| P0 (Critical) | As a user, I want the Agent to clearly display a human-readable summary before I sign, so I know exactly what I'm approving. | LLM-Powered Transaction Interpretation |
| P1 (High) | As a connected user, I want to type "Stake my SOL," so that the Agent automatically uses Marinade to generate the stake transaction and calculate the correct amount of SOL. | Balance-Aware Action |
| P1 (High) | As a user, I want the Agent to ask clarifying questions if my prompt is too vague (e.g., "Which token do you want to use?"), so I can correct the instruction. | Refinement/Ambiguity Handling |
The MVP will be delivered with the following feature set:
3.1. R/W (Read/Write) Engine
| Requirement | Description |
|---|---|
| FR1. Wallet Connection | Must use the Solana Wallet Adapter to connect major wallets (Phantom, Backpack). Must retrieve the user's Public Key (publicKey). |
| FR2. Balance Read | Upon connection, the Agent must fetch and maintain a list of all SPL Token and native SOL balances for the connected wallet. |
| FR3. Transaction Generation | Must construct a Serialized Transaction using the required on-chain instruction set (via Jupiter, Orca SDKs). |
| FR4. Transaction Broadcast | Must push the signed transaction to the Solana network (using a fast RPC, e.g., Helius/QuickNode). |
3.2. AI Agent Core
| Requirement | Description |
|---|---|
| FR5. Intent & Tool Selection | The LLM must accurately parse the user's natural language into a sequence of defined "Tools" (e.g., SwapTool, LPTool, StakeTool). |
| FR6. Token/Amount Parsing | The LLM must extract token symbols (e.g., USDC) and amounts from the input and validate them against FR2 (wallet balances). |
| FR7. Transaction Interpretation | Before signing, the LLM must generate a concise, step-by-step summary of the action and its consequences (e.g., "You are about to send $500 USDC to Orca..."). |
3.3. UX/UI Requirements
| Requirement | Description |
|---|---|
| FR8. Chat Interface | A simple, streaming chat window for user input and Agent output. |
| FR9. Pre-Execution Review | A dedicated panel must appear showing the FR7 summary and the estimated gas fee before the wallet signing prompt is triggered. |
| FR10. Status Updates | The UI must provide real-time updates: Processing Request, Generating Transaction, Waiting for Wallet Signature, Transaction Confirmed (with a link to Solscan). |
| Category | Requirement | Tools/Tech Stack |
|---|---|---|
| Frontend | Responsive chat interface | React/Next.js, Tailwind CSS |
| AI/LLM Core | Reasoning, Tool selection, Interpretation | LangChain or Solana Agent Kit (for tool definition), OpenAI API (or similar LLM) |
| Wallet | Wallet connection, signing | @solana/wallet-adapter-react-ui, useWallet() |
| On-Chain Data/Execution | Balances, Transaction construction/sending | @solana/web3.js, Jupiter SDK, Orca SDK |
| Infrastructure | Fast network access | High-speed RPC (e.g., Helius) for reliable transaction broadcasting. |
| Code | Open-Source submission | GitHub Repository, detailed README.md |
To maintain focus and deliver a complete, high-quality MVP, the following are out of scope:
This structured approach will guide your team, keep the scope manageable, and provide the judges with a clear measure of your project's success. Good luck!