Market Snapshot (Skill-Only, OpenClaw)
This skill is designed for OpenClaw/ClawHub bots that need a fast, low-friction market snapshot.
What This Skill Does
- - Calls a hosted market snapshot endpoint (
/skills/market-snapshot) with one or more token queries. - The API resolves tokens + fetches pricing/metadata server-side (backed by Jupiter Tokens V2 + Price V3).
- Emits a stable JSON object (no prose) so other agents/bots can parse it reliably.
What This Skill Will Not Do
- - It will not create or manage wallets.
- It will not request, store, or handle private keys / seed phrases.
- It will not execute swaps or provide "trade recommendations".
How To Use
When the user asks for prices, market snapshots, token metadata, or "what is X doing", run a snapshot.
Input formats supported:
- - Symbols:
SOL, USDC, INLINECODE3 - Names:
solana, INLINECODE5 - Mints: INLINECODE6
If multiple tokens are provided, resolve all of them and return a combined snapshot.
Data Sources (GET, No Headers Needed)
- INLINECODE7
Examples (copy/paste):
-
https://app.vecstack.com/api/skills/market-snapshot?q=SOL&source=openclaw
- - Multiple tokens (comma-separated, no spaces):
- INLINECODE9
Notes:
- -
web_fetch caches by URL. If the user explicitly needs "fresh right now" data, append a cache-buster query param like &_t=<unix> to the URL. - Do not invent values. If a fetch fails, keep
null fields and include an entry in warnings/errors.
Output Contract (Return JSON Only)
Return a single JSON object with this shape:
CODEBLOCK0
Field rules:
- -
as_of_unix: set to current Unix time when you finish assembling the response. - INLINECODE16 ,
mcap_usd, fdv_usd, and stats.* are populated from Tokens V2 search when present. - INLINECODE20 and
price_change_24h_pct are populated from Price V3 when present. - INLINECODE22 : non-fatal issues (missing price, ambiguous match, rate limits, etc).
- INLINECODE23 : fatal issues that prevented a snapshot (e.g., all sources failed).
Implementation Notes For OpenClaw
- - Prefer the
web_fetch tool for the endpoint, using extractMode=text so the body stays parseable as JSON. - If
web_fetch returns non-JSON content, retry once with a cache-buster (append &_t=<unix>). - Keep the final response strictly JSON.
市场快照（纯技能，OpenClaw）
此技能专为需要快速、低摩擦市场快照的OpenClaw/ClawHub机器人设计。
此技能的功能
- - 调用托管的市场快照端点（/skills/market-snapshot），支持一个或多个代币查询。
- API在服务端解析代币并获取定价/元数据（基于Jupiter Tokens V2 + Price V3）。
- 输出稳定的JSON对象（无文本描述），便于其他代理/机器人可靠解析。
此技能不会执行的操作
- - 不会创建或管理钱包。
- 不会请求、存储或处理私钥/助记词。
- 不会执行兑换或提供交易建议。
使用方法
当用户询问价格、市场快照、代币元数据或X在做什么时，运行快照。
支持的输入格式：
- - 符号：SOL、USDC、JUP
- 名称：solana、jupiter
- 铸币地址：So11111111111111111111111111111111111111112
如果提供多个代币，解析所有代币并返回合并后的快照。
数据源（GET请求，无需请求头）
- https://app.vecstack.com/api/skills/market-snapshot?q=
&source=openclaw
示例（可直接复制粘贴）：
- https://app.vecstack.com/api/skills/market-snapshot?q=SOL&source=openclaw
- https://app.vecstack.com/api/skills/market-snapshot?q=SOL,USDC,JUP&source=openclaw
注意：
- - webfetch按URL缓存。如果用户明确需要当前最新数据，请在URL后附加缓存清除参数，如&t=。
- 不要虚构数值。如果获取失败，保留null字段，并在warnings/errors中添加条目。
输出约定（仅返回JSON）
返回一个具有以下结构的JSON对象：
json
{
asofunix: 0,
provider: jupiter,
inputs: [SOL, USDC],
tokens: [
{
query: SOL,
mint: So11111111111111111111111111111111111111112,
symbol: SOL,
name: Wrapped SOL,
decimals: 9,
verified: true,
tags: [],
liquidity_usd: null,
mcap_usd: null,
fdv_usd: null,
usd_price: null,
pricechange24h_pct: null,
stats: {
5m: {
pricechangepct: null,
volume_usd: null,
buyvolumeusd: null,
sellvolumeusd: null
},
1h: {
pricechangepct: null,
volume_usd: null,
buyvolumeusd: null,
sellvolumeusd: null
},
24h: {
pricechangepct: null,
volume_usd: null,
buyvolumeusd: null,
sellvolumeusd: null
}
},
sources: {
tokensearchurl: null,
price_url: null
}
}
],
warnings: [],
errors: []
}
字段规则：
- - asofunix：在完成响应组装时设置为当前Unix时间。
- liquidityusd、mcapusd、fdvusd和stats.*：当存在时从Tokens V2搜索中填充。
- usdprice和pricechange24h_pct：当存在时从Price V3中填充。
- warnings：非致命问题（缺少价格、模糊匹配、速率限制等）。
- errors：阻止快照的致命问题（例如，所有数据源均失败）。
OpenClaw实现说明
- - 优先使用webfetch工具访问端点，设置extractMode=text以保持响应体可解析为JSON。
- 如果webfetch返回非JSON内容，使用缓存清除参数重试一次（追加&_t=）。
- 最终响应严格保持JSON格式。
