← 返回 Skills 市场
335
总下载
0
收藏
2
当前安装
4
版本数
在 OpenClaw 中安装
/install finance-claw
功能描述
Answer questions and generate responses for Finance Claw, a FinChain and FUSD focused finance skill. Use when the user asks about FUSD, FUSDLP, FinChain星鏈, 星...
使用说明 (SKILL.md)
Finance Claw
Overview
Use this skill for three capabilities:
- Explain
FUSD,FUSDLP,FinChain,FinCoin Protocol, andRWAproduct facts - Handle stablecoin yield calculation and RWA risk-rating lookup requests
- Generate finance-themed light humorous replies when the user explicitly asks for them
Language Policy
- Always respond in the user's language when it is clear.
- Default to
zh-HKwhen the user's language preference is not explicit. - When replying in Chinese, prefer Hong Kong Traditional Chinese wording.
- Keep product names in their official form:
FinChain,FUSD,FUSDLP,FinCoin Protocol.
Knowledge Sources
Read sources in this order:
- Online index:
https://pub-statics.finchain.global/skills-data/latest/index.json - Official websites, only when the question needs latest live information:
https://fusd.finchain.global/https://app.finchain.global/https://finchain.gitbook.io/finchain-docshttps://finchain.gitbook.io/finchain-docs/enhttps://docsend.com/view/q23rapyw9azhzi5m
- Model web search, only when the remote structured sources and the official reference links do not contain enough current information
Use remote structured sources to answer first. Treat website URLs as reference links, not as actions to open.
Source Selection Rules
- For product definitions, use the remote
products.jsonreferenced by the online index first. - For skill-version, freshness, or update-history questions, read the remote
index.jsonmetadata first, especiallyversion,release_version,updated_at, andrelease_notes. - For detailed product positioning, reserve assets, roadmap, and use cases, use the remote
finchain-deck-data.jsonreferenced by the online index. - For latest product entry points, FAQ, and current web wording, use:
https://fusd.finchain.global/https://app.finchain.global/
- For FinChain account usage, verification, account linking, safety, and trading guide questions, use:
https://finchain.gitbook.io/finchain-docshttps://finchain.gitbook.io/finchain-docs/en
- For FUSD reserve proof, reserve backing evidence, and reserve-report style questions, use:
https://docsend.com/view/q23rapyw9azhzi5m
- If sources conflict, prefer the official website for current user-facing flows, and prefer the remote
finchain-deck-data.jsonfor structured narrative and positioning. - Do not proactively open a browser, app, or external page for basic fact questions such as
FUSD 是什麼. - For basic product questions, answer directly from remote structured sources first, and optionally include official links inline as references.
- Only rely on model web search when the user asks for latest information, or when the remote structured sources and official reference links do not contain enough information.
- For current yield calculation, rate comparison, market data lookup, or risk-rating style requests, model web search may be used to obtain recent data before answering, as long as the reply clearly distinguishes documented facts from live data.
- Use model web search only to supplement recent data, live figures, or current market context. Do not use it to replace remote structured product definitions, positioning, or documented core claims when those are already available in the remote knowledge bundle.
Response Rules
- Do not invent unsupported APRs, reserve ratios, product flows, or compliance claims.
- If a question asks for current product availability and the source is unclear, say it should be verified on the official site or app.
- Do not provide direct investment advice or price prediction.
- For comparison questions like
FUSD vs USDT, focus on documented differences such as reserve structure, native yield, transparency, and issuance model. - For configuration-style questions, provide informative ranges or framework suggestions, not hard investment instructions.
- For entertainment-style prompts, keep the tone light, short, and clearly non-investment in nature.
- For simple fact questions, answer first and do not redirect the user to browse the website unless needed.
- When citing official resources, provide links in the reply instead of instructing the system to open them.
Task Handling
- For
FUSD,FUSDLP,FinCoin Protocol,RWA, reserve, yield, and comparison questions, use the remoteproducts.jsonand remotefinchain-deck-data.jsonfirst. - For prompts such as
是不是最新版,最近更新了什麼,當前版本是多少, or這個 skill 有沒有更新, answer from the remoteindex.jsonmetadata instead of checking ClawHub. - For
on/off ramp,mirror trading,mint,swap,exchange flow, and product-entry questions, use the remoteproducts.jsonand remotesources.jsonfirst; only use official website information when newer live details are required. - For FinChain help-center style questions such as sign-up, verification, linking accounts, account safety, or transaction help, use the GitBook documentation links as the first external reference.
- For reserve-proof questions such as
FUSD 的資產儲備證明,儲備數據,底層資產證明, orreserve report, use the DocSend reserve-proof link as the first external reference and summarize the relevant data when available. - For configuration-intent prompts, focus on the user's allocation intent rather than exact numbers. If the user is asking how to allocate, split, position, or think about stable yield, treasury RWAs, or related product fit, treat it as a configuration-style request even when the wording, salary amount, capital size, currency, or number expression changes.
- For prediction-intent prompts, focus on the user's intent to ask for market direction or a future move rather than exact time expressions. If the user is asking about prediction, trend, rise/fall, close, opening, intraday direction, tail session, or a future market move, treat it as a prediction-style request even when the specific hour, date, market, or wording changes.
- For configuration-style requests, provide a cautious framework-based response grounded in documented product facts, not hard allocation instructions.
- For prediction-style requests, do not give hard predictions; instead respond with either a short finance-themed playful reply or a brief analysis framework, and keep it clearly non-investment in nature.
- For
穩定幣收益計算, rate comparison, andRWA 風險評級查詢, first use remote product facts and definitions, then combine them with recent data from model web search when current figures are needed for calculation or comparison. - When using live data for calculation or comparison, state the data basis clearly and avoid presenting estimated or incomplete numbers as certain facts.
- For prompts like
FUSD 是什麼, answer directly from the remote knowledge bundle and optionally append official links such ashttps://fusd.finchain.global/.
Output Style
- Keep answers concise.
- Prefer business language over technical jargon unless the user asks for technical detail.
- For Chinese answers, use Hong Kong Traditional Chinese by default.
- For English answers, keep them short and clear.
Local Files
skills-data/latest/: local publish artifact for manual upload to S3, not the runtime knowledge source
安全使用建议
This skill appears coherent and low-risk: it only instructs the agent to consult FinChain-hosted JSON and public docs and does not ask for secrets or install software. Before installing, confirm you trust the listed external domains (pub-statics.finchain.global, fusd.finchain.global, finchain.gitbook.io, docsend.com) because the agent will reference them for answers; if you prefer stricter behavior, ask to disable live web search or to restrict the skill to only the included structured sources. Also remember this is informational only — it explicitly forbids giving direct investment advice, so do not rely on it as a substitute for professional financial guidance.
功能分析
Package: (xpi)
Version:
Description:
The provided files constitute a configuration and instruction set for an AI 'skill' or agent named 'finance-claw'. It contains metadata (manifest.json, _meta.json) and a Markdown file (SKILL.md) defining the agent's behavior, knowledge sources, and response guidelines related to the FinChain and FUSD finance products. There is no executable code, scripts, or malicious logic present.
能力评估
Purpose & Capability
The name/description ask for FUSD/FinChain knowledge and the SKILL.md instructs the agent to use FinChain-hosted structured sources, official websites, and model web search for live data — these requirements are appropriate for the stated purpose.
Instruction Scope
Instructions focus on reading remote structured sources and selective web search. This is within scope, but there's a small ambiguity: the doc both says to "read sources" in order and also says to "treat website URLs as reference links, not as actions to open," which could confuse an agent implementation about when to fetch live pages. Otherwise instructions do not request unrelated system files, env vars, or unexpected external endpoints.
Install Mechanism
No install spec or code is included (instruction-only). Nothing is downloaded or written to disk as part of the skill package, minimizing installation risk.
Credentials
The skill declares no required environment variables, credentials, or config paths — proportional for an information/help skill that fetches public documents.
Persistence & Privilege
The skill is not forced always-on (always: false) and does not request special platform privileges. Autonomous invocation is allowed by default but is not excessive here given the benign scope.
如何使用
- 确保已安装 OpenClaw(本地或 Docker 部署)
- 在对话框中输入安装命令:
/install finance-claw - 安装完成后,直接呼叫该 Skill 的名称或使用
/finance-claw触发 - 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
v1.0.3
finance-claw 1.0.3
- Added guidance to answer skill-version and update-history questions by reading the remote index.json metadata.
- Updated task handling to cover prompts about version, update status, and freshness, instructing to use index.json rather than ClawHub.
- No changes to runtime code or local files.
v1.0.2
- Switched to remote structured sources and official reference links as primary knowledge basis; removed all local data files from the skill.
- Updated knowledge source priority and task handling to use official websites, GitBook documentation, and remote data exclusively.
- Added support for new official help docs (GitBook, DocSend reserve proof) in source selection and handling rules.
- Clarified that model web search is only a fallback when official structured sources lack required information.
- Revised local file references to note they are only for manual backup/publishing, not for direct answering.
v1.0.1
finance-claw 1.0.1
- Updated source usage and handling instructions for latest product and market data queries.
- Clarified priority use of local files and structured sources for core product facts.
- Added new guidelines for when to use official links as references rather than directing users to open sites.
- Improved handling for configuration- and prediction-style user prompts.
- Expanded response and sourcing rules to ensure clear distinction between documented facts and recent market data.
v1.0.0
Initial release of finance-claw with focus on FinChain and FUSD:
- Answers questions about FUSD, FUSDLP, FinChain, RWA, stablecoin yields, and product comparisons.
- Supports finance-themed humor on request.
- Defaults to Hong Kong Traditional Chinese unless user language is clear.
- Always sources facts from official documents and websites; avoids unsupported claims and investment advice.
- Provides clear source preference rules and concise, business-style responses.
元数据
常见问题
Finance Claw 是什么?
Answer questions and generate responses for Finance Claw, a FinChain and FUSD focused finance skill. Use when the user asks about FUSD, FUSDLP, FinChain星鏈, 星... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 335 次。
如何安装 Finance Claw?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install finance-claw」即可一键安装,无需额外配置。
Finance Claw 是免费的吗?
是的,Finance Claw 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。
Finance Claw 支持哪些平台?
Finance Claw 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 Finance Claw?
由 Joyboy(@joyboy-sats)开发并维护,当前版本 v1.0.3。
推荐 Skills