菜单
Use for targeted EVM chain, account, transaction, RPC, explorer, and bridge data: chain name/ID, native symbol, RouteMesh, balances, token/NFT holdings/transfers, tx history, funding origin via Etherscan/Blockscout/Chainscout; Bungee, LayerZero, LI.FI, Socket enrichment.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
Use only when explicitly asked to archive/prune/compact/roll over checked tasks from TODO.md into `.ai/todos/TODO_UNTIL_YYYY_MM_DD.md`, leaving unchecked tasks.
Use only when explicitly invoked for Git commit workflows: stage intended changes, craft Conventional Prefix Format messages by default, Natural Language messages with --natural or configured repos, commit, and optionally --all, --staged, --deep, --close, or --push.
Use for dependency updates: update/bump deps, npm/pnpm/yarn/bun package upgrades, outdated checks, package.json updates, or taze.
Use for release versioning: bump/cut/tag a release, bump version, create a release, changelog updates, or version tagging.
Use for GitHub PR/issue/discussion workflows: create/update PRs or issues, post comments, start discussions; triggers include create/open PR, file/update issue, yeet.
Use to polish recently changed code: simplify for readability/maintainability and run a risk-profiled review that autonomously applies fixes. Default runs both passes; pass --simplify or --review for one. Covers code/PR review, audits, bug/security checks, reviewing diffs or changes, cleanup, refactoring, and reducing complexity.
| argument-hint | <chain-name-or-id> |
| disable-model-invocation | false |
| name | evm-atlas |
| user-invocable | true |
| description | Use for targeted EVM chain, account, transaction, RPC, explorer, and bridge data: chain name/ID, native symbol, RouteMesh, balances, token/NFT holdings/transfers, tx history, funding origin via Etherscan/Blockscout/Chainscout; Bungee, LayerZero, LI.FI, Socket enrichment. |
Targeted EVM chain atlas (chain name, chain ID, public RPCs, native currency symbol, default block explorer URL, RouteMesh support) and a router for reading on-chain data: resolve the chain, then dispatch balance, token, transfer, transaction, and first-funding queries to Etherscan (preferred) or Blockscout (fallback). For bridge transactions and cross-chain swaps, enrich explorer/RPC verification with Bungee, LayerZero, or LI.FI data only when the relevant chains are target chains.
Use this skill to resolve chain metadata before reading from an RPC, sending transactions, calling contracts, constructing chain-specific RPC URLs, or building explorer links to addresses, transactions, or blocks — and to query on-chain account data once the chain is resolved (see Querying On-Chain Data (Routing)). Also use it when the user mentions bridging, bridge tx, cross-chain swap, Bungee, LayerZero, LI.FI, Socket, Stargate, OFT, CCTP, or Aori, or when a transaction is inferred to be bridge-related.
Match chains by displayed name or numeric chain ID. ./references/target-mainnets.json is the skill's authoritative scope. If the requested chain is not listed there, do not use Etherscan, Blockscout, Bungee, LayerZero, LI.FI, RouteMesh, Chainlist, web search, or public RPCs to work around the scope. Ask the user to file a feature request in https://github.com/PaulRBerg/agent-skills.
Also normalize common code aliases from ./references/chain-aliases.json to their target mainnet rows.
When the user provides an EVM address and asks whether it was ever active on any chain, which chains it used, or whether it holds balances now across chains, read ./references/address-sweeps.md before the single-chain router below. The sweep scope is still only ./references/target-mainnets.json, not every provider-supported EVM chain.
If the prompt names a specific target chain, use Querying On-Chain Data (Routing) instead.
To read account data — native balance, token holdings, ERC-20/721/1155 transfers, transaction history, or first-funding — resolve the chain, then dispatch to the right explorer API. Do not default to Ethereum; infer the chain from the prompt (explicit chain mention, chain-specific tokens like POL→137 / ARB→42161, target-chain aliases). If ambiguous, ask. If the resolved chain is not a target chain, ask for a feature request instead of querying.
./references/target-mainnets.json and ./references/chain-aliases.json../references/etherscan-chains.md, follow ./references/etherscan-api.md (unified API V2, needs $ETHERSCAN_API_KEY)../references/blockscout-api.md.cast from the cli-cast skill, or curl). If the chain is not a target chain, ask for a feature request instead of querying.Paid-chain auto-fallback. Base (8453), Optimism (10), Avalanche (43114), and BNB Chain (56) are Etherscan-listed target chains, but their data endpoints require a paid Etherscan plan. If the target is one of these and ./scripts/etherscan-detect-plan.sh reports paid_chains=false (free tier), route to Blockscout instead. Etherscan stays the default for every other Etherscan-listed target chain.
When the user mentions bridging, bridge tx, cross-chain swap, or a bridge provider, or when a transaction looks bridge-related from logs, counterparties, calldata, or token movement, first confirm the known origin and destination chains are target chains. If either known chain is outside ./references/target-mainnets.json, ask for a feature request instead of continuing.
Then route to the bridge reference that matches the prompt or evidence:
./references/bungee-api.md../references/layerzero-bridging.md../references/lifi-bridging.md.Use bridge APIs as read-only enrichment sources alongside Etherscan, Blockscout, explorers, and RPC receipts. Do not treat bridge APIs as authoritative for on-chain execution by themselves; verify submitted transactions and terminal outcomes with explorer/RPC data whenever possible. Returned route steps, calldata, userSteps, transactionRequest, or signatures are for inspection only in this skill. Do not sign messages, submit signatures, execute user steps, or broadcast bridge transactions.
Use RouteMesh only when the routeMesh field in ./references/target-mainnets.json is true and the ROUTEMESH_API_KEY environment variable is available.
To verify current RouteMesh support, call GET https://lb.routeme.sh/chains; use https://lb2.routeme.sh/chains as the backup endpoint. Do not use https://rpc.routeme.sh/chains; the hostname may not resolve even though it appears in RouteMesh's OpenAPI spec.
Construct the RouteMesh RPC URL as:
https://lb.routeme.sh/rpc/CHAIN_ID/ROUTEMESH_API_KEY
Replace CHAIN_ID with the numeric chain ID and ROUTEMESH_API_KEY with the value of the ROUTEMESH_API_KEY environment variable. If routeMesh is false or ROUTEMESH_API_KEY is not available, use the chain's primary public RPC first, then the listed fallback RPCs in order.
Public RPCs are best-effort. Before relying on one for data fetches or contract calls, verify it with eth_chainId. If the primary endpoint fails, try the fallback endpoints for that chain from ./references/target-fallback-rpcs.json in order. If a chain has no fallback row, only the primary public RPC is listed.
Only use RPCs for target chains. For any chain not in ./references/target-mainnets.json, ask the user to file a feature request in https://github.com/PaulRBerg/agent-skills.
The explorerUrl field in ./references/target-mainnets.json is the base URL of the chain's canonical block explorer. Append the path patterns from ./references/explorer-paths.json to build links.
Etherscan and Etherscan-stack explorers (Arbiscan, Basescan, BscScan, Polygonscan, Optimism Etherscan, Lineascan, Snowscan, Blastscan, Berascan, Uniscan, Gnosisscan, abscan.org) all follow this scheme. Most Blockscout-based and chain-native target explorers accept the same segments, but conventions can drift — verify against the explorer UI when in doubt.
Explorer URL presence, Etherscan-style paths, or an Etherscan-stack explorer name do not imply Etherscan API V2 support. When API coverage matters, use ./references/etherscan-chains.md. Etherscan's live https://api.etherscan.io/v2/chainlist endpoint may contain additional provider-supported chains, but those are outside this skill unless they appear in ./references/target-mainnets.json. Treat this JSON as chain metadata for RPC and explorer-link construction only.
OP Mainnet pre-regenesis history is not available through current explorer/RPC routes. For OP Mainnet (10) queries before the final regenesis on 2021-11-11, read ./references/optimism-pre-2021-11-11.md before using Etherscan, Blockscout, or public RPC results.
Ronin (app.roninchain.com) does not follow the Etherscan path scheme. Verify against the explorer UI before constructing a Ronin link.
Ronin (2020) collides with a non-target Chainscout registry entry. Do not use Chainscout metadata for Ronin; use Ronin's listed explorer/RPC or another target-aware source.
Target mainnet metadata lives in ./references/target-mainnets.json. The slug field is the Sablier SDK chain slug (sablier package ~/sablier/sdk, src/evm/chains/specs.ts, resolved as meta.slug ?? key). Arbitrum Nova, Celo, Fantom, IoTeX, and Zora are not defined in the SDK; their slugs follow the same convention (lowercase, hyphenated name).
Target fallback RPC metadata lives in ./references/target-fallback-rpcs.json. Use each chain's fallbackPublicRpcs values in order after the primaryPublicRpc from ./references/target-mainnets.json fails verification.