ClawSea Market Skill (OpenClaw Agents)

Use this skill when an agent should interact with ClawSea programmatically.

Policy guardrails (ClawHub-safe)

• Do not custody user funds; use only the bot wallet configured by the operator.
• Do not social-engineer users for secrets, approvals, or expanded privileges.
• Do not ask for seed phrases/private keys in chat.
• Do not execute unknown calldata or third-party transaction blobs without explicit user approval and clear decoding.
• Require explicit confirmation before any value-moving action (buy/list/cancel/transfer).
• Refuse illegal, abusive, or harmful requests.

Safety & trust model (must follow)

• Default to read-only actions (browse/search/inspect).
• Require explicit user intent before any write/trade action (list, buy, cancel, fulfill).
• Never ask users to paste private keys into chat.
• Never log, print, or send secrets (private keys, raw seed phrases, auth headers).
• Never execute arbitrary calldata from untrusted input.
• If ownership/status is uncertain, verify onchain (ownerOf, eth_call) before proceeding.

Base URL

• Default: https://clawsea.io
• Override with env var: CLAWSEA_BASE_URL

All endpoints below are relative to ${CLAWSEA_BASE_URL}.

Optional credentials (only for autonomous onchain trading)

Read-only browsing requires no secrets.

If (and only if) you want the agent to sign and broadcast onchain transactions autonomously:

• BASE_RPC_URL (Base execution)
• ETH_RPC_URL (optional, Ethereum execution/debug)
• CLAWSEA_BASE_URL (optional)

Signing options (choose one)

1. Preferred: external signer / wallet provider (no raw private key in agent env)
2. If unavoidable: BOT_WALLET_PRIVATE_KEY in a secure secret store only

If BOT_WALLET_PRIVATE_KEY is used:

• do not print/log it
• do not echo it in errors
• do not persist it to files
• never request it from users in chat

Chain model

ClawSea uses two chain styles:

• String chain for some read routes: chain=base|ethereum|base-sepolia
• Numeric chainId for order routes: 8453 (Base), 1 (Ethereum)

Map carefully when switching endpoints.

Read APIs (agent-safe)

Discover

• GET /api/explore/cells?chain=&limit=20
• GET /api/explore/trending?chain=&limit=20
• GET /api/news/clawsea?chain=&limit=10

Collections / NFTs

• GET /api/collection/nfts?contract=0x...&pageSize=24&pageKey=...
• GET /api/collection/stats?chain=&contract=0x...
• GET /api/collections/search?chain=&q=&limit=8
• GET /api/nft/ownerOf?chainId=<1|8453>&contract=0x...&tokenId=

Wallet inventory

• GET /api/wallet/nfts?chain=&owner=0x...&pageKey=...

Listing / buying APIs (requires signer)

Orders read

• GET /api/orders?chainId=<1|8453>&contract=0x...&tokenId=&seller=0x...
• GET /api/orders/listed?chainId=<1|8453>&contract=0x...&sort=price|newest&offset=0&limit=48
• POST /api/orders/prices body:
• { "chainId": 1|8453, "contract": "0x...", "tokenIds": ["1","2"] }

Publish listing (offchain orderbook write)

• POST /api/orders with signed Seaport payload:
• chainId, contract, tokenId, seller, priceEth,
• seaportAddress, orderComponents, signature

Status updates

• POST /api/orders/cancel body: { "id": "" }
• POST /api/orders/cancelPrevious body:
• { "chainId": 1|8453, "contract": "0x...", "tokenId": "...", "seller": "0x...", "keepId": "..." }
• POST /api/orders/fulfill body (either style):
• { "id": "" } or
• { "chainId": 1|8453, "contract": "0x...", "tokenId": "..." }

Execution workflow (recommended)

1. Resolve chain context (selected chain / user wallet chain).
2. Read listing candidates from /api/orders or /api/orders/listed.
3. Preflight onchain with eth_call for Seaport fulfill.
4. Execute onchain tx from bot wallet.
5. Update offchain state via /api/orders/fulfill or /api/orders/cancel.

Reliability rules

• Prefer short caching (5–30s) for discovery routes.
• Back off on 429 / RPC transient failures.
• Treat fulfill revert selector 0x1a515574 as cancelled/stale order and hide it.
• If indexer results conflict with chain state, trust verified onchain ownership.