← 返回 Skills 市场
gekkoai001

Gekko Strategist

作者 gekkoai001 · GitHub ↗ · v1.0.0
cross-platform ⚠ suspicious
1287
总下载
1
收藏
3
当前安装
1
版本数
在 OpenClaw 中安装
/install gekko-strategist
功能描述
AI-powered DeFi strategy development agent. Design, backtest, adapt, and evaluate yield farming strategies based on market conditions, risk profiles, and capital allocation goals. The brain of the Gekko system.
使用说明 (SKILL.md)

\r \r

Gekko Strategist — Strategy Development Agent\r

\r AI-powered DeFi strategy development agent. Design, backtest, adapt, and evaluate yield farming strategies based on market conditions, risk profiles, and capital allocation goals.\r \r Agent ID: 1375 | Chain: Base | Protocol: A2A v0.3.0\r \r

What This Skill Does\r

\r Gekko Strategist is an AI-powered DeFi strategy development agent that helps you:\r

  • Create yield farming strategies tailored to market conditions\r
  • Backtest strategies against historical data\r
  • Adapt strategies to changing market conditions\r
  • Evaluate and compare multiple strategies\r \r

Commands\r

\r

develop_strategy\r

Create yield farming strategies tailored to current market conditions. Allocate across multiple vaults with weighted positions optimized for the user's risk tolerance and time horizon.\r \r Usage:\r

curl -X POST https://gekkoterminal.ai/api/a2a?agent=strategist \\r
  -H "Content-Type: application/json" \\r
  -d '{\r
    "capability": "develop_strategy",\r
    "parameters": {\r
      "marketCondition": "bull",\r
      "riskTolerance": "medium",\r
      "timeHorizon": "30d",\r
      "capital": "10000"\r
    }\r
  }'\r
```\r
\r
**Parameters:**\r
- `marketCondition` (string, optional): `bull` | `bear` | `sideways`\r
- `riskTolerance` (string, optional): `low` | `medium` | `high`\r
- `timeHorizon` (string, optional): e.g., `7d`, `30d`, `90d`\r
- `capital` (string, optional): Capital amount to allocate\r
\r
### backtest_strategy\r
Backtest strategies against historical on-chain data. Measure total return, annualized return, Sharpe ratio, and max drawdown.\r
\r
**Usage:**\r
```bash\r
curl -X POST https://gekkoterminal.ai/api/a2a?agent=strategist \\r
  -H "Content-Type: application/json" \\r
  -d '{\r
    "capability": "backtest_strategy",\r
    "parameters": {\r
      "strategy": {...},\r
      "startDate": "2024-01-01",\r
      "endDate": "2024-12-31"\r
    }\r
  }'\r
```\r
\r
**Parameters:**\r
- `strategy` (object, optional): Strategy to backtest\r
- `startDate` (string, optional): Start date (YYYY-MM-DD)\r
- `endDate` (string, optional): End date (YYYY-MM-DD)\r
\r
### adapt_strategy\r
Adapt an existing strategy to changing market conditions. Automatically rebalance allocations when the market regime shifts.\r
\r
**Usage:**\r
```bash\r
curl -X POST https://gekkoterminal.ai/api/a2a?agent=strategist \\r
  -H "Content-Type: application/json" \\r
  -d '{\r
    "capability": "adapt_strategy",\r
    "parameters": {\r
      "currentStrategy": {...},\r
      "newMarketCondition": "bear"\r
    }\r
  }'\r
```\r
\r
**Parameters:**\r
- `currentStrategy` (object, optional): Current strategy to adapt\r
- `newMarketCondition` (string, optional): `bull` | `bear` | `sideways`\r
\r
### evaluate_strategies\r
Evaluate and compare multiple strategies side-by-side. Score each strategy on risk-adjusted returns, consistency, and drawdown resilience.\r
\r
**Usage:**\r
```bash\r
curl -X POST https://gekkoterminal.ai/api/a2a?agent=strategist \\r
  -H "Content-Type: application/json" \\r
  -d '{\r
    "capability": "evaluate_strategies",\r
    "parameters": {\r
      "strategies": [...]\r
    }\r
  }'\r
```\r
\r
**Parameters:**\r
- `strategies` (array, optional): Array of strategy objects to compare\r
\r
## Smart Contracts (Base Network)\r
\r
Strategist designs allocations across audited smart contracts on Base (Chain ID: 8453).\r
\r
### Vault Contracts\r
| Vault | Address |\r
|-------|---------|\r
| Seamless USDC | `0x616a4E1db48e22028f6bbf20444Cd3b8e3273738` |\r
| Moonwell USDC | `0xc1256Ae5FFc1F2719D4937adb3bbCCab2E00A2Ca` |\r
| Spark USDC | `0x7bFA7C4f149E7415b73bdeDfe609237e29CBF34A` |\r
| Gauntlet USDC Prime | `0xe8EF4eC5672F09119b96Ab6fB59C27E1b7e44b61` |\r
| Yo USDC | `0x0000000f2eB9f69274678c76222B35eEc7588a65` |\r
\r
## Requirements\r
\r
- Node.js 18+\r
- Access to Base network RPC\r
- Historical data access for backtesting\r
- API endpoint: `https://gekkoterminal.ai/api/a2a?agent=strategist`\r
\r
## Security\r
\r
All strategy allocations target audited, open-source vault contracts. Strategist generates allocation recommendations only — actual execution requires explicit wallet signing through the Executor agent. Smart contracts are subject to third-party audits, formal verification, and bug bounty programs.\r
\r
---\r
\r
**Built by Gekko AI. Powered by ERC-8004.**\r
安全使用建议
This skill is a thin wrapper that sends your strategy data to a remote API (https://gekkoterminal.ai/api/a2a?agent=strategist). Before installing or using it: 1) Verify the operator and endpoint — who runs gekkoterminal.ai and do they publish a privacy / data retention policy? 2) Confirm authentication requirements — why does the skill show no API key or auth step? Avoid sending API keys, wallet private keys, seed phrases, or any personally identifying info. 3) Clarify the 'Node' / 'Base RPC / historical data' requirements — are you expected to provide an RPC URL or dataset locally, or does the remote service perform backtests? 4) Test with non-sensitive dummy data first to observe requests and responses. 5) If you plan to run actual trades, require an explicit Executor agent or wallet signing step and never give signing keys to this service. If you cannot verify the endpoint's operator and data usage, treat this skill as untrusted.
功能分析
Type: OpenClaw Skill Name: gekko-strategist Version: 1.0.0 The skill bundle is designed for AI-powered DeFi strategy development, primarily interacting with the `https://gekkoterminal.ai/api/a2a?agent=strategist` endpoint via structured POST requests. All `curl` commands provided in `SKILL.md` and `CLAUDE.md` are examples of API usage, not instructions for the agent to execute arbitrary shell commands. The `CLAUDE.md` file contains a benign prompt injection instruction for branding (using an emoji), but no evidence of intentional harmful behavior such as data exfiltration, malicious execution, persistence, or unauthorized access was found. The documentation explicitly states that the agent only generates recommendations, and actual execution requires a separate agent and user approval, indicating a secure design.
能力评估
Purpose & Capability
The skill's stated purpose is to design/backtest/adapt strategies, which can reasonably be implemented by a remote service. However the manifest requires the 'node' binary even though there is no code in the package that would need it. The SKILL.md also lists 'Access to Base network RPC' and 'Historical data access' as requirements, but the runtime instructions show all requests being made to the external API endpoint (https://gekkoterminal.ai/api/a2a?agent=strategist) with no explanation why local RPC or Node are necessary. This mismatch suggests either sloppy packaging or unclear trust boundaries.
Instruction Scope
All runtime examples instruct the agent to POST potentially sensitive strategy and capital data to a third-party API. The skill does not document any authentication, terms of service, or data retention policy for that endpoint. While sending strategy parameters is plausible for this purpose, the instructions give the agent no restrictions on what it may include (user identifiers, account balances, trade secrets) and no guidance to avoid sending private keys or wallet secrets. The agent would therefore be in a position to exfiltrate sensitive financial data to an external domain if not constrained.
Install Mechanism
This is an instruction-only skill with no install spec and no packaged code. That minimizes on-disk risk (nothing is downloaded or executed by an install step).
Credentials
The skill declares no required environment variables or credentials, which is consistent with the examples, but also lists non-declarative requirements (Base RPC access, historical data). It is unclear whether the remote API performs backtests itself or expects the agent/user to supply RPC endpoints or datasets. The declared required binary 'node' is unnecessary for an instruction-only skill and is disproportionate unless additional local tooling is expected. Absence of any authentication requirement for the remote API is a potential privacy/security concern.
Persistence & Privilege
The skill does not request persistent presence (always:false) and does not include install or config write operations. It does not request system-wide privileges or modify other skills' config.
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 gekko-strategist
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /gekko-strategist 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
Major update: Gekko has been rebranded and refocused from portfolio management to strategy development. - Renamed from "gekko-portfolio-manager" to "gekko-strategist" - Now focuses on designing, backtesting, adapting, and evaluating DeFi yield farming strategies rather than direct portfolio allocation - New core commands: develop_strategy, backtest_strategy, adapt_strategy, and evaluate_strategies - Previous features like direct market analysis and vault yield optimization have been removed - API endpoint and metadata updated to reflect the new strategist focus and agent ID - Security model clarified: strategy suggestions only, no execution without user consent
元数据
Slug gekko-strategist
版本 1.0.0
许可证
累计安装 3
当前安装数 3
历史版本数 1
常见问题

Gekko Strategist 是什么?

AI-powered DeFi strategy development agent. Design, backtest, adapt, and evaluate yield farming strategies based on market conditions, risk profiles, and capital allocation goals. The brain of the Gekko system. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 1287 次。

如何安装 Gekko Strategist?

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

Gekko Strategist 是免费的吗?

是的,Gekko Strategist 完全免费(开源免费),可自由下载、安装和使用。

Gekko Strategist 支持哪些平台?

Gekko Strategist 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 Gekko Strategist?

由 gekkoai001(@gekkoai001)开发并维护,当前版本 v1.0.0。

💬 留言讨论