Install
$ agentstack add mcp-vo1ganin-crypto-claude-skills ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ● Network access Used
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ● Environment & secrets Used
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Crypto Claude Skills
> Production-ready Claude Code skills for on-chain data APIs — Dune Analytics, Solscan Pro, Nansen, and universal Solana RPC (Helius / QuickNode / any provider).
[](LICENSE) [](#skills) [](https://claude.com/claude-code) [](AGENTS.md)
Each skill enforces credit budget rules, prefers batch/parsed APIs over raw loops, and ships with ready-to-run async Python examples — built from real-world on-chain research (copytrade analysis, bot reverse-engineering, wallet profiling).
Table of Contents
- [Skills](#skills)
- [Install](#install)
- [Configure API keys](#configure-api-keys)
- [Cross-agent support](#cross-agent-support)
- [Design principles](#design-principles)
- [Repository layout](#repository-layout)
- [Contributing](#contributing)
- [License](#license)
Skills
| Skill | Standalone repo | Use for | Signature feature | |-------|-----------------|---------|-------------------| | [dune](skills/dune) | dune-skill | SQL on 100+ chains (Dune Analytics) | 500 / 700 credit budget gates, FREE ↔ PAID key rotation | | [solscan](skills/solscan) | solscan-skill | Solana wallet/token/NFT data (Solscan Pro v2) | Export endpoints 15× cheaper than pagination, multi-endpoints 50× cheaper | | [nansen](skills/nansen) | nansen-skill | Smart Money + wallet profiling on 37 chains | premium_labels: true 30× cost trap, live credit balance headers | | [solana-rpc](skills/solana-rpc) | solana-rpc-skill | Raw Solana RPC (Helius / QuickNode / Ankr / …) | JSON-RPC array batching, DAS over getProgramAccounts, Helius Enhanced Tx | | [pumpfun](skills/pumpfun) | pumpfun-skill | pump.fun bonding curves + PumpPortal API | WebSocket streaming (single-conn rule), Lightning vs Local trading, sniper/copytrade examples | | [dexscreener](skills/dexscreener) | dexscreener-skill | Current DEX pair data across 20+ chains | Free/no-auth API, batch 30 tokens per call, 300 rpm fast tier | | [mev-bundles](skills/mev-bundles) | mev-bundles-skill | MEV bundles + bribes: what/how to find/how to identify | Jito + 8 other relays (Bloxroute, Nozomi, Astralane, BlockRazor, Stellium, Falcon, Flashblocks, 1node), sandwich detection | | [coinmarketcap](skills/coinmarketcap) | coinmarketcap-skill | CoinMarketCap Pro API (prices, OHLCV, F&G, global metrics) | Credit-aware batching (100 IDs per call), symbol→ID caching, live credit_count logging |
Each skill folder contains SKILL.md, topic references/*.md, and runnable references/examples/*.py. Each skill is also available as a standalone public GitHub repo — install just the one you need from its own repo.
Install
# Clone
git clone https://github.com/Vo1ganin/crypto-claude-skills.git
cd crypto-claude-skills
# Install all eight skills to Claude Code
mkdir -p ~/.claude/skills
for s in dune solscan nansen solana-rpc pumpfun dexscreener mev-bundles coinmarketcap; do
cp -R skills/$s ~/.claude/skills/
done
Restart Claude Code. Skills trigger automatically based on your prompt (e.g. "analyze this Solana wallet" → Solscan or solana-rpc fires).
For one-at-a-time install or to run Python examples standalone, see [INSTALL.md](INSTALL.md).
Configure API keys
cp .env.example .env
# edit .env with your keys
set -a; source .env; set +a
| Variable | For | |----------|-----| | DUNE_API_KEY_FREE | Dune (default) | | DUNE_API_KEY_PAID | Dune (bulk exports, optional) | | SOLSCAN_API_KEY | Solscan Pro v2 | | NANSEN_API_KEY | Nansen | | SOLANA_RPC_URL | Any Solana RPC provider (Helius, QuickNode, …) |
Full key reference with examples in [.env.example](.env.example).
Cross-agent support
Primary target is Claude Code — but the content works across other AI coding agents too.
| Agent | What to do | How it's used | |-------|-----------|---------------| | Claude Code | cp -R skills/ ~/.claude/skills/ | Auto-triggers on prompt keywords | | Claude Agent SDK | Load SKILL.md programmatically | Your own Claude-based bot | | OpenAI Codex / Codex CLI | Reads [AGENTS.md](AGENTS.md) automatically on project load | Cross-skill instructions | | Cursor / Windsurf | Copy SKILL.md into .cursor/rules/ | Per-project rules | | MCP-aware clients (Claude, Cursor, ChatGPT, Codex, OpenCode) | Configure Dune + Solscan MCP servers per .env.example | Tool calls work natively | | Any LLM via API | Include SKILL.md as system prompt | Self-contained instructions |
Note: skills/*/references/examples/*.py are provider-agnostic — they work with plain Python, no assistant needed.
See [AGENTS.md](AGENTS.md) for the full agents.md-spec instruction set.
Design principles
💰 Credits are real money
Every skill enforces budget thresholds before expensive calls. Dune: 500 warn / 700 block per operation. Nansen: 50 warn / 200 block. Skills announce estimated cost and reason when approaching a threshold and refuse to proceed above the hard limit without explicit user approval.
🤖 Scripts for batches, MCP for exploration
- ≤ 10 API calls → MCP tools or inline
curl - \> 10 calls of the same shape → async Python script with
asyncio.Semaphore, resume-safe JSONL output, live rate-limit monitoring via response headers
MCP-in-a-loop is a documented antipattern — see each skill's references/*-patterns.md.
🚀 Parsed > raw, batch > loop
- Helius Enhanced Transactions >
getSignaturesForAddress+ N ×getTransaction - DAS
getAssetsByOwner>getTokenAccountsByOwner+ metadata - Solscan
transaction/detail/multi(50 sigs / 100 CU) > 50 ×transaction/detail(5000 CU) - Dune
VALUESbatch clauses > many small queries - JSON-RPC array requests > N sequential HTTP POSTs
🔁 Multi-key rotation (safely)
Dune and Nansen skills support multiple keys per tier. On 402/429 they rotate FREE keys automatically. Escalation to PAID key requires explicit user consent every time — never silent.
Repository layout
crypto-claude-skills/
├── README.md ← you are here
├── INSTALL.md ← step-by-step install & troubleshooting
├── CONTRIBUTING.md
├── CHANGELOG.md
├── LICENSE
├── .env.example ← key template (never commit .env)
├── .gitignore
├── docs/ ← per-provider documentation corpus
│ ├── dune/
│ ├── solscan/
│ ├── nansen/ ← includes llms-full.txt (official LLM-ready docs)
│ └── quicknode/ ← Solana RPC + Helius llms.txt
└── skills/ ← four Claude Code skills (copy to ~/.claude/skills/)
├── dune/
├── solscan/
├── nansen/
└── solana-rpc/
Contributing
See [CONTRIBUTING.md](CONTRIBUTING.md). Particularly welcome:
- New provider extensions (Triton, Ankr, Chainstack specifics)
- Additional
references/examples/*.pyscripts for common use cases - Cost/limit table updates when providers change pricing
License
[MIT](LICENSE)
Source & license
This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: Vo1ganin
- Source: Vo1ganin/crypto-claude-skills
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.