Token导航 LogoToken导航

Agent Skill · 研究检索

simplify-budget-expense-tracker(简化预算支出跟踪器)

simplify-budget-expense-tracker 用于查找、检索和筛选相关信息,适合在 OpenClaw 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

simplify-budget-expense-tracker 是研究检索类 Agent Skill,可用宿主为 OpenClaw,通过ClawHub获取,权限为接触敏感数据;许可证为 MIT-0,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 15,427。

总安装 7,858周安装 321GitHub Stars 1下载 2,517核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
MIT-0
版本
1.1.6

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:simplify-budget-expense-tracker(简化预算支出跟踪器)
来源仓库:https://github.com/serdarsalim/simplify-budget-expense-tracker
安装命令:
openclaw skills install simplify-budget-expense-tracker
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 OpenClaw 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

ClawHubOpenClaw
openclaw skills install simplify-budget-expense-tracker

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

simplify-budget-expense-tracker 用于查找、检索和筛选相关信息,适合在 OpenClaw 中需要根据关键词快速定位候选结果时使用。主要功能是在 Google 表格中记录、更新和删除费用与收入,并回答只读问题。

通过 openclaw skills install 命令从 ClawHub 安装,需确认权限范围。建议在使用前检查维护状态及是否会触发 Google Sheets API 调用。适用宿主包括 OpenClaw,接入前应确认版本、权限和运行环境要求。

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 OpenClaw;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 openclaw skills install simplify-budget-expense-tracker;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要接触敏感数据,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

simplify-budget-expense-tracker 怎么安装?

在终端执行 openclaw skills install simplify-budget-expense-tracker。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

simplify-budget-expense-tracker 支持哪些 AI 工具?

公开资料标注支持 OpenClaw。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

simplify-budget-expense-tracker 需要什么权限?

权限级别是接触敏感数据。该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

simplify-budget-expense-tracker 是免费的吗?

来源标注的许可证是 MIT-0。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

simplify-budget-expense-tracker 用的人多吗?

公开数据是累计安装 7,858 次,近一周安装 321 次,下载 2,517 次,GitHub 1 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

simplify-budget-expense-tracker 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 专注 OpenClaw 一种宿主,触发方式和使用习惯不用来回切换。
  • 许可证明确为 MIT-0,商用和分发边界好判断。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 需要接触敏感数据,装到常用环境前要先评估影响面。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 OpenClaw,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。
  • 需要许可证清楚、并且有第三方扫描结果的团队项目。

不太适合

  • 环境里有大量密钥、凭证,不方便给第三方 Skill 读取的人。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
simplify-budget-expense-tracker

简化预算支出跟踪器 · 本页

7,8581接触敏感数据MIT-0通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

name
simplify-budget
description
Log, find, update, and delete expenses and income in the Simplify Budget Google Sheet, and answer read-only recurring schedule questions. NEVER use sessions_spawn or ACP — ONLY use the exec tool to run bash scripts. Categories are hardcoded in this skill — NEVER invent category names, ALWAYS pick from the exact list provided. Income uses name, account, source, and notes. For edits/deletes, find rows first, then mutate by transaction id. Amounts are always stored in the configured tracker currency. Just do it.
version
1.2.0
user-invocable
true
metadata
openclaw
primaryEnv
GOOGLE_SA_FILE
emoji
💸
requires
env
bins

Simplify Budget Tracker

CRITICAL EXECUTION RULE: You MUST use the exec tool to run the bash scripts below. Do NOT call sessions_spawn. Do NOT create ACP sessions. These are standalone shell scripts. Resolve script paths relative to this skill directory and run the resulting absolute path with exec.

🔴 SCRIPT-FIRST RULE — NON-NEGOTIABLE

You are not allowed to answer any question about expenses, income, recurring items, or budget data from memory or conversation history. Ever. Not once.

Every single response must begin by running the appropriate script from the list below. The script output is your only source of truth. Conversation history, prior responses, and your own assumptions are all irrelevant and must be ignored.

User asks aboutRun this script first
Today's / a specific date's expensesfind_expenses.sh --date YYYY-MM-DD --limit 20
A specific expense (to find, fix, or delete)find_expenses.sh "<query>" 10
Monthly spending totalsfind_summary.sh --month YYYY-MM
Today's / a specific date's incomefind_income.sh --date YYYY-MM-DD --limit 20
A specific income entryfind_income.sh "<query>" 10
Recurring itemsfind_recurring.sh --month YYYY-MM or find_recurring.sh --query "<name>"
Logging a new expenselog.sh "<user message>"
Logging new incomewrite_income.sh

If you have not run the script yet, you have not started your response yet. Run the script. Then respond conversationally based on what it returns.

⛔ CATEGORY RULES — READ BEFORE ANYTHING ELSE

WRONG (never do this):

"Here are suggested categories: Fuel/Transport, Food & Drink, Medical/Pharmacy. Please confirm."

CORRECT (always do this):

"Here's what I'll log: Shell → Transport 🚙, Baker's Cottage → Groceries 🍎, Klinik → Medical 🩺. Shall I go ahead?"

Rules that must never be broken:

  1. Category names MUST come from the exact list below — never invent names like "Fuel/Transport" or "Food & Drink"
  2. Always confirm with the user before writing — relay the question field from preview_expense.sh verbatim
  3. NEVER say "your choice" or ask the user to pick from a list — the script picks and you relay its suggestion
  4. DO NOT call get_categories.sh — the full category list is hardcoded right here:
FormulaName
=zategory1Housing 🏡
=zategory2Transport 🚙
=zategory3Groceries 🍎
=zategory4Dining Out 🍕
=zategory5Personal Care ❤️
=zategory6Shopping 🛍️
=zategory7Utilities 💡
=zategory8Fun 🎬
=zategory9Business 💻️
=zategory10Other ❓
=zategory11Donation 🕌
=zategory12Childcare 🐣
=zategory13Travel ✈️
=zategory14Zakat 🌟
=zategory15Debt Payment 💸
=zategory16Fitness 💪
=zategory17Family Support 🏘️
=zategory18Taxes 💵
=zategory19Maintenance 🧰
=zategory20Painting 🎨
=zategory21TestGround 🤖
=zategory22Learning 📚
=zategory23Sports 🏀
=zategory24Pet 🐶
=zategory25Gifts 🎁
=zategory26Special Occasions 🥰
=zategory27Dress 👚
=zategory28Hobby 🪂
=zategory29Insurance 🛡️
=zategory30Medical 🩺

These are the ONLY valid categories. Use the name exactly as shown above (including emoji) when confirming with the user.

When to use

  • The user mentions spending money, making a purchase, paying for something
  • The user says things like "I spent X on Y", "X euros for coffee", "paid X for Y", "log X"
  • The user asks to fix, correct, undo, or update the last expense entry
  • The user asks to search for or identify an existing expense
  • The user asks what they've spent or asks about their budget
  • The user wants to add an expense to their budget tracker
  • The user wants to log income, salary, withdrawal, remittance, sale proceeds, or any other incoming money
  • The user asks what recurring payments are due this month, when something recurring is due, or what subscriptions are upcoming

Configuration

Required environment variables:

  • GOOGLE_SA_FILE — absolute path to the Google service account JSON file
  • SPREADSHEET_ID — the Simplify Budget spreadsheet ID
  • TRACKER_CURRENCY — the base currency code for the tracker (for example EUR)

Optional environment variables:

  • TRACKER_CURRENCY_SYMBOL — display symbol for the base currency (for example €)

⛔ NEVER ASK THE USER TO REFORMAT

If the user says "log a pencil for 10 euro under business" — log it. Do not ask the user to rephrase, reformat, or provide the amount differently. Extract what you need from their message and run the script. This applies to every possible phrasing. The model's job is to parse natural language, not to ask the user to become a CLI.

If a script returns an error, report the actual error message. Never invent an error or silently fail.

Currency Rules

  • The configured TRACKER_CURRENCY is the system of record for stored amounts.
  • If the user gives an amount without a currency, assume it is already in TRACKER_CURRENCY.
  • If the user gives an explicit foreign currency, keep that currency in the amount argument you pass to the script, for example "50 MYR" or "12 USD".
  • The scripts handle natural language amounts directly. Pass these as-is:

- "10 euro" → --amount "10 euro" ✅ - "€10" → --amount "€10" ✅ - "10 euros" → --amount "10 euros" ✅ - "50 ringgit" → --amount "50 ringgit" ✅ - "12 dollars" → --amount "12 dollars" ✅ - "10" → --amount "10" ✅

  • The scripts fetch a live ECB FX rate, convert into TRACKER_CURRENCY, and store the converted amount in the sheet.
  • When conversion happens, the scripts append an [auto-fx] audit line to notes with the original amount, converted amount, rate, and rate date.
  • Never read the sheet to discover the base currency during normal operation. Use the configured environment instead.

Category Matching Rules

  1. Use the category list in the ⛔ CATEGORY RULES section above — never call get_categories.sh during normal operation
  2. The script picks the category and generates the confirmation question — relay it verbatim, do not pick or rephrase.
  3. Pass the plain English category name to the script (e.g. "Dining Out", "Business", "Transport"). The script resolves it to the correct formula internally. Never construct =zategory{N} yourself.
  4. Disambiguation hints:

- Coffee or café → Dining Out - Netflix/Spotify/streaming → Utilities - Concert, bar, night out → Fun - Gym or fitness class → Fitness - Doctor, pharmacy, clinic → Medical - Clothes, shoes, fashion → Dress - Electronics, gadgets, general retail → Shopping - Birthday gift or present → Gifts - Flight, hotel, vacation → Travel - Daily commute, taxi, fuel → Transport - Books, courses, online learning → Learning - Zakat → Zakat - Test entry → TestGround (only if user says it's a test)

Workflows

Log a new expense

When the user mentions spending money, buying something, paying for something, or asks to log an expense — follow this flow:

Step 1 — Preview:

bash <skill_dir>/scripts/log.sh --preview "<user's message, word-for-word>"

Pass the raw message. Do not reformat. Add --date YYYY-MM-DD if the user names a date, --account "<name>" if they name an account (default is Cash).

The script parses amount, description, and category. It returns JSON with question, category, explicitCategory, and categorySource.

Step 2 — Confirm or write directly:

  • If explicitCategory is true: skip confirmation, go straight to Step 3.
  • If categorySource is "builtin" or "learned": relay the question field verbatim and wait for the user's reply.
  • Otherwise: relay the question field (it includes a "best guess" note) and wait.

When the user replies:

  • "yes" → proceed with the proposed category
  • A category name (e.g. "Transport") → use that category instead
  • "no" / "cancel" → stop

Step 3 — Write:

bash <skill_dir>/scripts/log.sh "<user's message>" --category "<confirmed category>"

The script writes to the sheet and outputs a REPLY: line. Send everything after REPLY: verbatim. Do not paraphrase.

Examples:

  • User: i bought a pencil for 10 euro under business category

→ preview detects explicit category → write directly (no confirmation needed)

  • User: spent 23 on mcdonalds

→ preview suggests Dining Out (builtin match) → ask "Log mcdonalds under Dining Out 🍕?" → on yes, write

  • User: €12 coffee

→ preview suggests Dining Out → confirm → write

Rules:

  1. Never edit or reformat the user's message. Pass it verbatim.
  2. Never ask the user to rephrase. The script handles any phrasing.
  3. For multiple expenses in one message (e.g. "10 on coffee and 5 on parking"), run the preview → confirm → write flow once per expense, sequentially.
  4. Duplicate check: if the write step outputs DUPLICATE_FOUND:, tell the user and ask if they want to log anyway. If yes, re-run with --skip-duplicate-check.
  5. Relay REPLY: lines verbatim — do not paraphrase or confirm from memory.

Log an expense from a receipt image

When the user uploads a receipt image or asks you to log a receipt:

  1. Read the receipt visually and extract the likely:

- merchant - final charged total or grand total - transaction date if visible - broad purchase type

  1. Default to exactly one expense row per receipt.
  2. Do NOT create one expense per line item unless the user explicitly asks to split the receipt.
  3. Prefer the actual charged total, final total, or grand total.

- Do not use subtotal unless that is the only total shown. - Do not sum visible item lines if a final total is already present.

  1. Use the merchant name or a short summary as the description.
  2. Best-match the whole receipt into one real category using the Active Categories table above.
  3. If the receipt is materially mixed, such as fuel plus snacks:

- default to one expense using the dominant purpose - ask one short clarification only if the split is important or the category is genuinely ambiguous

  1. Then run the normal expense write flow.

Examples:

  • grocery receipt with 10 items -> one Groceries expense using the grand total
  • restaurant receipt with food and tax lines -> one Dining Out expense using the final total
  • petrol station receipt with mostly fuel and a small snack -> one Transport expense by default unless the user asks to split it

Find or inspect expenses

When the user wants to inspect, fix, or delete an expense, resolve it from the sheet first:

  1. Build a short natural-language query from the user's message.
  2. Search the sheet:
   bash <skill_dir>/scripts/find_expenses.sh "<query>" 10
  1. Matching MUST consider both description and notes.
  2. If one clear match exists, proceed. If multiple plausible matches exist, ask one short disambiguation question.

Log a new income

When the user provides income (amount + name, with optional date/account/source/notes):

  1. Extract from the user's message:

- amount — required. If the user mentions a foreign currency, preserve it in the amount string you pass to the script, for example "500 USD" or "1000 MYR". If they give no currency, pass a plain number. - name — required. This is the income title, e.g. Salary, BMW Sale, Etoro withdrawal - date — in YYYY-MM-DD format. Default to today if not specified. - account — default to Other if not specified. - source — default to Other if not specified. Use the user’s wording when it is clear, e.g. Salary, Capital Gains, Remittance, Crypto, MTS. - notes — optional supporting context.

  1. Write the income:
   bash <skill_dir>/scripts/write_income.sh "<amount_or_amount_with_currency>" "<name>" "<YYYY-MM-DD>" "<account>" "<source>" "<notes>"
  1. Confirm to the user in a concise way:

"✅ Logged income [name] — [amount] into [account] from [source] on [date]" Include notes only when present.

Find or inspect income

When the user wants to inspect, fix, or delete income, resolve it from the sheet first:

  1. Build a short natural-language query from the user's message.
  2. Search the sheet:
   bash <skill_dir>/scripts/find_income.sh "<query>" 10
  1. Matching MUST consider name, source, and notes.
  2. If one clear match exists, proceed. If multiple plausible matches exist, ask one short disambiguation question.

Check today's or a specific day's expenses

Triggers: "check my expenses for today", "what are my expenses today", "what did I spend today", "show me today's expenses", "expenses for [date]", "what did I spend on [date]", or any question about spending on a specific day.

Always run find_expenses.sh with a date filter. This is not optional.

bash <skill_dir>/scripts/find_expenses.sh --date <YYYY-MM-DD> --limit 20
  • Use the script output to answer. Never use conversation history or memory.
  • If the script returns entries, list them: [description] — [amount] under [category] ([account])
  • Only say "no expenses for today" if the script returns an empty array [].

find_summary.sh is for month-level questions only. Never use it for today/yesterday/specific-date queries.

Read monthly totals

Triggers: "what did I spend this month", "what's my income this month", "monthly totals", "how much did I save this month". These are month-level questions with no specific date.

bash <skill_dir>/scripts/find_summary.sh --month YYYY-MM

Rules:

  1. This is read-only.
  2. Dontedit is the source of truth for monthly totals.
  3. Never use find_summary.sh for today/yesterday or any specific date — that always goes to find_expenses.sh.
  4. Respond concisely with income, spending, and savings when available.

Inspect recurring schedule

When the user asks read-only recurring questions like:

  • what is due this month
  • when is capcut due
  • what subscriptions are due next

Use the recurring query script. Do NOT write anything into Expenses or Income.

Examples:

bash <skill_dir>/scripts/find_recurring.sh --month 2026-03
bash <skill_dir>/scripts/find_recurring.sh --query "CapCut" --date 2026-03-28
bash <skill_dir>/scripts/find_recurring.sh --query "CapCut" --mode next --date 2026-03-28

Rules:

  1. This is read-only. Never materialize recurring rows into the ledgers.
  2. The script calculates cycles from Recurring using the same recurrence rules as the existing Apps Script logic.
  3. For month-style questions, return the current-month cycle entries.
  4. For when is X due, prefer --mode next.
  5. Respond concisely with the due date, amount, account, and whether it is expense or income.

Fix or delete recurring items

When the user wants to change or delete a recurring item in the Recurring tab:

  1. Resolve the target recurring row using find_recurring.sh.
  2. If one clear match exists, mutate the Recurring row itself. Never write anything into Expenses or Income for this task.
  3. To update a recurring row, use __KEEP__ for unchanged fields and __CLEAR__ for optional end date / notes / source:
   bash <skill_dir>/scripts/update_recurring.sh "<recurring_id>" "<YYYY-MM-DD_or___KEEP__>" "<name_or___KEEP__>" "<category_or___KEEP__>" "<expense_or_income_or___KEEP__>" "<Monthly_or_Quarterly_or_Yearly_or___KEEP__>" "<amount_or___KEEP__>" "<account_or___KEEP__>" "<YYYY-MM-DD_or___KEEP___or___CLEAR__>" "<notes_or___KEEP___or___CLEAR__>" "<source_or___KEEP___or___CLEAR__>"
  1. To delete a recurring row, clear the row by recurring id:
   bash <skill_dir>/scripts/delete_recurring.sh "<recurring_id>"
  1. Confirm concisely what changed.

Add a recurring item

When the user wants to add a recurring expense or recurring income to the Recurring tab:

  1. Extract:

- start_date in YYYY-MM-DD - name - category must use the live active category list; never invent a category - type as expense or income - frequency as Monthly, Quarterly, or Yearly - amount - optional account, end_date, notes, source

  1. Write it with:
   bash <skill_dir>/scripts/write_recurring.sh "<YYYY-MM-DD>" "<name>" "<category>" "<expense_or_income>" "<Monthly_or_Quarterly_or_Yearly>" "<amount>" "<account>" "<YYYY-MM-DD_optional_end_date>" "<notes>" "<source>"
  1. This must reuse the first empty row in Recurring starting from row 6, matching the SB_LIVE hole-reuse behavior.
  2. Recurring categories follow SB_LIVE rules:

- expense recurring items must store a =zategory<stableId> formula — use the Active Categories table above to resolve it - recurring income is the only case that may store the literal Income 💵

  1. Never ask the user to pick a category unless they explicitly want to choose. Best-match into an existing category.
  2. Confirm concisely.

Fix or correct income

When the user wants to change amount, name, date, account, source, or notes for an income row:

  1. Resolve the target income from the sheet using find_income.sh. Do NOT trust chat memory as the source of truth.
  2. Infer the correction from their message. If the new amount is in a foreign currency, preserve that currency in the amount argument you pass to the script.
  3. Run the update. Use __KEEP__ for unchanged fields and __CLEAR__ to blank notes:
   bash <skill_dir>/scripts/update_income.sh "<transaction_id>" "<amount_or_amount_with_currency_or___KEEP__>" "<name_or___KEEP__>" "<YYYY-MM-DD_or___KEEP__>" "<account_or___KEEP__>" "<source_or___KEEP__>" "<notes_or___KEEP___or___CLEAR__>"
  1. Confirm: "✅ Updated income — now [name] — [amount] into [account] from [source]"

Delete income

If the user asks to undo or delete an income entry:

  1. Resolve the target income from the sheet using find_income.sh.
  2. If there is one clear match, clear the row by transaction id:
   bash <skill_dir>/scripts/delete_income.sh "<transaction_id>"
  1. Confirm that the income row was cleared.
  2. Never delete sheet rows. Clear the existing row contents instead.

Fix or correct the last expense

When the user says things like "fix that", "that was wrong", "change the amount", "put that under X instead", "it was 4 not 5", "actually that was yesterday", "actually I bought that on [date]", "wrong date":

  1. Run find_expenses.sh to resolve the target from the sheet. Do not use chat memory.
  2. Infer the correction (amount, category, description, date, account, notes).
  3. Run the update. Use __KEEP__ for unchanged fields, __CLEAR__ to blank notes. Pass category as plain English name:
   bash <skill_dir>/scripts/update_expense.sh --id "<transaction_id>" --amount "<amount_or___KEEP__>" --description "<description_or___KEEP__>" --category "<category_name_or___KEEP__>" --date "<YYYY-MM-DD_or___KEEP__>" --account "<account_or___KEEP__>" --notes "<notes_or___KEEP___or___CLEAR__>"
  1. Relay the REPLY: line from the script output word for word.

Undo / delete last expense

If the user asks to undo or delete an entry:

  1. Run find_expenses.sh to resolve the target from the sheet.
  2. If one clear match, delete it:
   bash <skill_dir>/scripts/delete_expense.sh "<transaction_id>"
  1. Relay the REPLY: line from the script output word for word.
  2. Never delete sheet rows — the script clears the row contents.

Rules

  • Pass category as plain English to scripts — never construct =zategory{N} yourself
  • Do NOT call get_categories.sh at runtime — the category table is hardcoded in the scripts
  • For income, use the explicit name, account, source, and notes fields.
  • Relay REPLY: lines from scripts verbatim — do not paraphrase or confirm from memory
  • Never include transaction IDs or internal record IDs in any response
  • Never use conversation history as evidence of what is or isn't in the sheet — always check the sheet
  • Always relay what the script confirmed — the user should never have to guess if it worked
  • If a script returns an error, tell the user clearly and do not silently retry
  • Default date is always today in the user's local timezone
  • Default account is always "Revolut" unless the user specifies otherwise
  • Default income account is always "Other" unless the user specifies otherwise
  • Default income source is always "Other" unless the user specifies otherwise
  • Amounts are always stored as plain numbers in TRACKER_CURRENCY
  • If a foreign currency is provided, keep it in the script input and let the script convert it into TRACKER_CURRENCY
  • Notes are a first-class field. Search them, preserve them on update unless changed, and clear them on delete.
  • The sheet is always the source of truth — for reads, edits, deletes, and duplicate checks.
  • Recurring schedule questions are read-only. Never create expense or income rows just because the user asked what is due.
  • The 🤖 label on written rows identifies bot-added entries in the sheet

平台分布

OpenClaw

79.71%
按下载量换算2,006

安全审计

VirusTotal

通过

ClawScan

通过

Static analysis

通过

权限和风险

接触敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills