Best Claude Skills for Crypto (2026)
Most "best skills" lists are copied README summaries. This one isn't: we installed all 34 crypto skills in our catalog, ran each against a baseline and scored four criteria. 7 passed cleanly, average score 8.6/10. The one to start with: Vly Money.
Ranked by SkillProof Score · how we score
- 1
Vly Money
Builds vly.money deep links for crypto sends and X402 paywalled-content redirects -- no execution, just correct URLs.
From our test: Fed it 'send 0.01 USDC to <Solana address> on Solana' and it produced the exact deterministic vly.money URL per its own table; fed it an invalid pair (vUSD on Solana, which the table restricts to the…
- 2
Quicknode Skill
Quicknode's own reference skill covering RPC, Streams, SQL Explorer, x402/MPP agent payments, CLI, MCP, and more.
From our test: Asked it to wire up a keyless, wallet-paid x402 RPC call with no Quicknode API key. Baseline (general x402 knowledge only) got the protocol right but had to guess at Quicknode's actual gateway domain…
- 3
Hyperliquid (ChainGPT)
Live read-only Hyperliquid mainnet data — markets, orderbook, account state, funding — via the ChainGPT MCP server.
From our test: Built and ran the real mcp-server over stdio against live Hyperliquid mainnet endpoints — chaingpt_hl_funding's auto-annualized BTC rate (10.95%) matched an independent manual curl-plus-arithmetic…
- 4
Smart Contract Vulnerability Auditor
4-phase Solidity vulnerability scan across 36 exploit classes (Solidity only).
From our test: Ran it against a 4-function test vault seeded with 3 bugs: it caught the reentrancy in withdraw() (external call before balance update), the missing access control on setOwner(), and — only because…
- 5
Dune Sim Onchain Analytics
Dune Sim vs Dune Analytics SQL decision framework for wallet/token intelligence — guidance only, no bundled scripts.
From our test: Correctly routes 'watch this wallet live' to Sim's realtime endpoints and 'analyze 90 days of swap volume' to Dune Analytics SQL, with an explicit fact-vs-inference caveat on wallet labels — but it's…
- 6
Wallet Security Review
Structured 8-section wallet security checklist: keys, approvals, signing, phishing, opsec, multisig, incident response.
From our test: Fed it a MetaMask-with-photographed-seed-phrase-and-unlimited-approvals scenario and got back a structured 'Critical risk' verdict flagging the permit-signature exploit risk a generic answer missed…
- 7
Cosmos Vulnerability Scanner
Spawns parallel Cosmos SDK / CosmWasm scanners for chain-halt and fund-loss bugs across 54 documented patterns.
From our test: Fed it a synthetic Cosmos rewards module with a map-iteration bug and an unvalidated msg_server payout; the skill's own pattern docs caught both plus a third bug a generic review missed…
- 8
Drift Protocol SDK Guide
Reference guide and code patterns for the Drift Protocol perp/spot/vault SDK on Solana.
From our test: Cross-checked the documented npm packages against the live registry -- @drift-labs/sdk and @coral-xyz/anchor are real, actively-published packages (unlike a sibling skill's broken pinned version)…
- 9
Clone Contracts
Pulls verified Solidity source from Etherscan V2, resolves proxies, rewrites @-imports to relative paths.
From our test: Ran the core pipeline live on synthetic Etherscan-style JSON: it correctly split a multi-file source, rewrote an `@openzeppelin/...` import to the right relative `../lib/...` path, and laid out…
- 10
Arbitrum DApp Skill
Exact cargo-stylus / Foundry / nitro-devnode commands for scaffolding, testing, and deploying Arbitrum Stylus + Solidity dApps.
From our test: Every command is copy-paste exact (cargo stylus new/check/deploy/export-abi, devnode port 8547, chain-id 412346) and it proactively surfaces two non-obvious footguns a generic-knowledge baseline…
How this ranking works
Each skill gets a fresh install in Claude Code, three trigger phrasings, and the same task run with and without it. Scores: installation (5), trigger reliability (5), output vs baseline (10), documentation honesty (5) — normalized to a 10-point SkillProof Score. Full rubric on the methodology page. Queue-verdict skills sit at the bottom until testing completes.
New here? Start with how to install Claude skills — or grab the whole setup in one command with a curated pack ($10).
On a different AI tool? These tested instructions adapt: see our conversion guides for Cursor rules, Copilot instructions, Windsurf and Custom GPTs.