← 返回 Skills 市场
eranp-orbs

AI agent crypto trading — gasless limit, DCA, stop-loss & take-profit

作者 Eranp-orbs · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ⚠ suspicious
15
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install agent-crypto-trading
功能描述
AI agent crypto trading. Gasless limit, DCA, stop-loss & take-profit across 8 EVM chains.
使用说明 (SKILL.md)

Spot Advanced Swap Orders

Give your AI agent autonomous crypto trading capabilities — gasless limit orders, DCA, stop-loss, and take-profit across 8 EVM chains. Use this skill when the agent needs to turn user intent into a final Spot order payload on a supported EVM chain. It covers order-shape selection, param normalization, typed-data population, approval guidance, signing, submission, query, and cancellation. This bundle is instruction-only: build everything locally from the bundled markdown and JSON assets, then submit only the final signed payload. Execution remains decentralized, non-custodial, oracle-protected, immutable, audited, and battle-tested onchain.

Supported Chains

  1. Ethereum - 1 - adapter 0xC1bB4d5071Fe7109ae2D67AE05826A3fe9116cfc
  2. BNB Chain - 56 - adapter 0x67Feba015c968c76cCB2EEabf197b4578640BE2C
  3. Polygon - 137 - adapter 0x75A3d70Fa6d054d31C896b9Cf8AB06b1c1B829B8
  4. Sonic - 146 - adapter 0xD87ee28806bc0060789C4789F123647f4Df25A6C
  5. Base - 8453 - adapter 0xc64d6E64A713EfbbCcB14413479c56461F9c0b77
  6. Arbitrum One - 42161 - adapter 0x6F1002141Fcb5d3A3aA8b12A49e6e7DCE5661ae9
  7. Avalanche - 43114 - adapter 0xc64d6E64A713EfbbCcB14413479c56461F9c0b77
  8. Linea - 59144 - adapter 0x55E4da2cd634729064bEb294EC682Dc94f5c3f24

Relay

  1. Submit signed orders with POST https://agents-sink.orbs.network/orders/new.
  2. Query orders with GET https://agents-sink.orbs.network/orders; see references/lifecycle.md for filters, polling, and cancellation follow-up.

Workflow

  1. Read references/quickstart.md for the minimum end-to-end flow.
  2. Use references/params.md to map user intent into params, defaults, validation, and order-shape fields.
  3. Use references/sign.md to fill the template, handle approval, sign, and submit.
  4. Use references/lifecycle.md for relay query semantics, status polling, and cancellation.
  5. Use references/examples.md only when the final relay payload shape is still unclear.
  6. Use assets/token-addressbook.md only for optional token alias lookup on supported chains.
  7. Use assets/repermit.template.json as the canonical typed-data shape.
  8. Treat ## Supported Chains as the authoritative source for chain support and per-chain adapters.
  9. Treat ## Relay as the authoritative relay endpoint list.

Guardrails

  1. ## Supported Chains is authoritative for chain support and per-chain adapters.
  2. ## Relay is authoritative for relay endpoints.
  3. assets/token-addressbook.md is a convenience alias list only. It does not expand chain support or override explicit user-provided addresses.
  4. This skill is instruction-only. Do not fetch or execute external helper code.
  5. Normalize params with references/params.md before touching the template.
  6. Replace only the \x3C...> placeholders in assets/repermit.template.json. Keep the fixed protocol fields already in the template unchanged.
  7. Default approval guidance is exact approve(..., input.maxAmount). Standing maxUint256 approval is opt-in convenience for repeat use, not the default suggestion.
  8. Submit only the final signed payload as described in references/sign.md.

Agent Contract

  1. Turn the user request into a params JSON object using references/params.md.
  2. Normalize params locally, including defaults, rounding, and order-shape fields.
  3. Confirm chainId is listed in ## Supported Chains, then populate assets/repermit.template.json from the normalized params and replace \x3CADAPTER> with that chain's listed adapter.
  4. Handle approval, signing, and submission exactly as described in references/sign.md, and forward the returned signature unchanged.
  5. Query and cancel exactly as described in references/lifecycle.md.
安全使用建议
Only install/use this if you understand the Spot protocol and are comfortable with an agent preparing crypto orders. Before signing anything, verify the chain, token addresses, spend amount, recipient, slippage, deadline, spender/verifying contract, and relay endpoint. Prefer exact approvals, avoid standing max approvals unless you deliberately choose them, and do not give the agent unattended access to a wallet or private key.
功能分析
Type: OpenClaw Skill Name: agent-crypto-trading Version: 1.0.0 The skill bundle provides a legitimate integration for the Orbs Spot protocol, enabling AI agents to facilitate non-custodial crypto trading (limit orders, DCA, stop-loss) across multiple EVM chains. It uses EIP-712 signatures for gasless execution via a dedicated relay service (agents-sink.orbs.network) and includes comprehensive documentation (SKILL.md, references/*.md) with explicit security guardrails and defensive warnings. No indicators of data exfiltration, malicious execution, or prompt injection were found; the workflow is transparent and aligned with standard DeFi practices.
能力标签
cryptorequires-walletrequires-sensitive-credentials
能力评估
Purpose & Capability
The trading, approval, signing, submission, query, and cancellation capabilities are aligned with the stated crypto-trading purpose, but they are high-impact financial actions involving real wallet assets.
Instruction Scope
The workflow describes autonomous order construction and exact handling of approval, signing, and submission, but it does not require a final human confirmation of chain, token, amount, recipient, slippage, deadline, and contract addresses before signing/submission.
Install Mechanism
There is no install spec and no executable code; the risk comes from the instructions rather than local code execution. Registry source/homepage information is limited, so users should verify provenance independently.
Credentials
Wallet signing and token approval are proportionate to a trading skill, but they cross a sensitive permission boundary and can mutate onchain balances through an external protocol and relay.
Persistence & Privilege
The instructions tell the agent to persist and reuse exact populated typed data and signatures after ambiguous relay failures, which may retain reusable financial authorization material without clear storage or retention limits.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install agent-crypto-trading
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /agent-crypto-trading 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.0
- Initial release of agent-crypto-trading skill. - Enables autonomous gasless limit, DCA, stop-loss, and take-profit orders on 8 EVM chains. - Provides complete workflow and relay endpoints for instruction-only crypto trading. - Includes clear mapping from user intent to spot order payloads, with param normalization, signing, and submission guidance. - Supports relay querying, order cancellation, and authoritative per-chain and adapter definitions.
元数据
Slug agent-crypto-trading
版本 1.0.0
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 1
常见问题

AI agent crypto trading — gasless limit, DCA, stop-loss & take-profit 是什么?

AI agent crypto trading. Gasless limit, DCA, stop-loss & take-profit across 8 EVM chains. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 15 次。

如何安装 AI agent crypto trading — gasless limit, DCA, stop-loss & take-profit?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install agent-crypto-trading」即可一键安装,无需额外配置。

AI agent crypto trading — gasless limit, DCA, stop-loss & take-profit 是免费的吗?

是的,AI agent crypto trading — gasless limit, DCA, stop-loss & take-profit 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

AI agent crypto trading — gasless limit, DCA, stop-loss & take-profit 支持哪些平台?

AI agent crypto trading — gasless limit, DCA, stop-loss & take-profit 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 AI agent crypto trading — gasless limit, DCA, stop-loss & take-profit?

由 Eranp-orbs(@eranp-orbs)开发并维护,当前版本 v1.0.0。

💬 留言讨论