惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

H
Hackread – Cybersecurity News, Data Breaches, AI and More
宝玉的分享
宝玉的分享
月光博客
月光博客
爱范儿
爱范儿
阮一峰的网络日志
阮一峰的网络日志
酷 壳 – CoolShell
酷 壳 – CoolShell
Recent Announcements
Recent Announcements
A
About on SuperTechFans
T
The Blog of Author Tim Ferriss
博客园 - 叶小钗
U
Unit 42
aimingoo的专栏
aimingoo的专栏
Y
Y Combinator Blog
Martin Fowler
Martin Fowler
N
Netflix TechBlog - Medium
博客园 - 司徒正美
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
云风的 BLOG
云风的 BLOG
M
MIT News - Artificial intelligence
大猫的无限游戏
大猫的无限游戏
J
Java Code Geeks
V
Visual Studio Blog
腾讯CDC
IT之家
IT之家

DEV Community

Authentication Security Deep Dive: From Brute Force to Salted Hashing (With Java Examples) Why AI Systems Don’t Fail — They Drift Spilling beans for how i learn for exam😁"Reinforcement Learning Cheat Sheet" I Replaced Chrome with Safari for AI Browser Automation. Here's What Broke (and What Finally Worked) How Python Borrows Other People's Work The $40 Architecture: Processing 1 Billion API Requests with 99.99% Uptime Vibe Coding: A Workflow Guide (From Zero to SaaS) Most webhook security guides protect the wrong side. The scary part is delivery. Headless CMS for TanStack Start: Build a Blog with Cosmic EU Age Verification App "Hacked in 2 Minutes" — What Actually Happened Comfy Cloud’s delete function does not actually remove files Running AI Models on GPU Cloud Servers: A Beginner Guide Event-driven media intelligence with AWS Step Functions and Bedrock I scored 500 AI prompts across 8 quality dimensions — here's what broke How to Call Google Gemini API from Next.js (Free Tier, No Backend Needed) The Portal Protocol: Reclaiming Human Connection in the Age of AI How to Fix Your Team's Scattered Knowledge Problem With a Self-Hosted Forum Intro to tc Cloud Functors: A Graph-First Mental Model for the Modern Cloud Designing Multi-Tenant Backends With Both Ownership and Team Access I Built a Neumorphic CSS Library with 77+ Components — Here's What I Learned PostgreSQL Performance Optimization: Why Connection Pooling Is Critical at Scale Cómo construí un SaaS multi-rubro para gestionar expensas en Argentina con FastAPI + Vue 3 🚀 I Built an Ethical Hacking Scanner Tool – Open Source Project I Replaced /usage and /context in Claude Code With a Single Statusline A Pythonic Way to Handle Emails (IMAP/SMTP) with Auto-Discovery and AI-Ready Design I Collected 8.9 Million Polymarket Price Points — Here's What I Found About How Markets Really Move EcoTrack AI — Carbon Footprint Tracker & Dashboard Everyone's Using AI. No One Agrees How. 5 self-hosted ebook managers worth trying in 2026 Building Your First AI Agent with LangChain: From Chatbot to Autonomous Assistant
Paste a wallet, get a personal airdrop verdict — and call...
Weston G · 2026-05-12 · via DEV Community

Weston G

Paste a wallet, get a personal airdrop verdict — and call the same logic from any LLM

Two surfaces, one rule registry: a browser tool and an MCP tool that share the exact same 14 hand-verified on-chain rules and return the exact same verdict shape.

The problem

Airdrop directories are full of "you might be eligible for 312 airdrops!" claims. They match on chain ("you've used Ethereum → you qualify for every Ethereum airdrop"), which is useless. The real eligibility lives on each project's contract or off-chain criteria.

So I built a tool that does the opposite: a tiny number of hand-verified rules, each tied to a specific entry, evaluated against the wallet you paste. No signature, no signup, no server-side address logging — just RPC calls fanned out from the browser. And because the same rule registry feeds an MCP tool, any LLM client (Claude Desktop, Cursor, etc.) can call it directly without scraping the site.

The browser tool

Live at web3-discover.vercel.app/tools/eligibility.

Paste an EVM and/or Solana address. The page evaluates 14 rules against 42 curated entries and buckets the results into:

  • ✅ Likely eligible — rule passes (e.g. eth_getTransactionCount on Linea ≥ 1, or you hold ≥ 1 PENDLE)
  • 🟡 Keep farming — partial: you've started but not crossed the threshold (e.g. 2 Arbitrum tx when 5 are needed for Reya / Ostium)
  • ⏭ Not relevant — rule fails outright
  • 🔍 Manual check — entry lives on a chain we can't reach from the browser (Monad, MegaETH, Hyperliquid, Sui-based) or has off-chain criteria (KYC, Discord, Galxe quests)

Try vitalik.eth + a public Solana wallet — resolves in 2-5 seconds, 9 eligible at the time of writing.

How the browser side stays cheap

Every entry-page CTA deep-links to /tools/eligibility?focus=<slug>. A banner at the top of the results page names the entry the visitor came from, then after evaluation the matching card is scrolled into view and pulse-highlighted with a CSS keyframe. This costs zero extra requests — the focus logic is pure DOM-after-render.

Total RPC requests = one batched JSON-RPC POST per EVM chain (so 6 max for an EVM address) plus per-mint Solana calls. Each batch bundles eth_getTransactionCount + every eth_call(balanceOf) we need for that chain, so adding more rules on Ethereum costs zero additional requests.

The MCP tool

Same logic, exposed at https://web3-discover.vercel.app/api/mcp as the fourth tool of the web3-discover MCP server.

curl -s -X POST https://web3-discover.vercel.app/api/mcp \
  -H 'content-type: application/json' \
  -d '{
    "jsonrpc": "2.0",
    "id": 7,
    "method": "tools/call",
    "params": {
      "name": "check_eligibility",
      "arguments": { "evm": "0xd8dA6BF26964aF9D7eEd9e03E53415D37aA96045" }
    }
  }'

Enter fullscreen mode Exit fullscreen mode

Returns the same verdict shape — eligible[], partial[], manual[], unavailable[], skip[] — but as machine-readable JSON each LLM client can render however it likes:

{
  "inputs": { "evm": "0xd8dA…6045", "sol": null },
  "rulesEvaluated": 14,
  "totalTracked": 42,
  "counts": {
    "eligible": 9,
    "partial": 0,
    "skip": 5,
    "manual": 28,
    "unavailable": 0
  },
  "eligible": [
    {
      "slug": "base-coinbase-l2",
      "project": "Base",
      "chain": "Base (Coinbase L2)",
      "verdictDetail": "70 tx on Base (≥ 1 required)",
      "officialUrl": "https://base.org"
    }
  ]
}

Enter fullscreen mode Exit fullscreen mode

Drop-in for Claude Desktop / Cursor

{
  "mcpServers": {
    "web3-discover": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://web3-discover.vercel.app/api/mcp"]
    }
  }
}

Enter fullscreen mode Exit fullscreen mode

Then ask: "What airdrops am I eligible for? My wallet is 0x…" and the model calls check_eligibility directly. No scraping, no flaky HTML parsing.

The one-file pattern that keeps both surfaces honest

Two surfaces means two ways for the rules to drift. The fix is one source of truth:

src/data/
  eligibility-rules.json   # source of truth
  eligibility-rules.ts     # typed wrapper for the Astro page

Enter fullscreen mode Exit fullscreen mode

And during prebuild:

// scripts/build-airdrops-snapshot.mjs
fs.copyFileSync(
  path.join(ROOT, 'src/data/eligibility-rules.json'),
  path.join(ROOT, 'api/_eligibility-rules.json'),
);

Enter fullscreen mode Exit fullscreen mode

The MCP serverless function reads the JSON at cold start; the Astro page imports the typed .ts wrapper. Same data, two consumers, zero drift.

Why this matters as a wedge

Most directory sites optimise for "discover something new". This optimises for the moment after you've already heard of a project: "do I, this wallet I'm holding, actually need to act on this?" That question doesn't compete with newsletters or Twitter — it complements them.

And exposing the same logic over MCP means the directory gets pulled into LLM chats as the user is making decisions, not after the fact. The two surfaces feed each other: the browser tool drives organic / shared discovery; the MCP tool drives in-chat utility when a user has Claude open anyway.

Try it

Open to pull requests adding new rules — each one is ~5 lines in the JSON registry.