论文审查
- 作者仓库星标 110
- 作者更新于 实时读取
- 作者仓库 botstreet
- 领域
- 通用
- 兼容 Agent
-
- Claude Code
- Cursor
- Cline
- Codex
- Windsurf
- Gemini CLI
- +20
- 信任分
- 88 / 100 · 社区维护
- 作者 / 版本 / 许可
- @upstash · 未声明 license
- Token 消耗评级
- 低消耗
- 接入复杂程度
- 需简单配置
- 是否需要外部 API Key
- 不需要
- 兼容的系统
- macOS · Linux · Windows
- 底层运行要求
- 无特殊要求
- 文件与系统权限
-
- 只读
- 允许写入 / 修改
- Shell 执行
- 网络行为
- 仅限本地
- 安装命令数
- 26 条
档案由构建时根据 SKILL.md 与安装命令自动衍生,可能与作者实际意图存在差异。
需要注意: 未限定 allowed-tools,默认拥有全部工具权限。
---
name: trade
description: Execute the daily virtual trading process. Research markets, make trades, save snapshot, write d…
category: 通用
runtime: 无特殊运行时
---
# trade 输出预览
## PART A: 任务判断
- 适用问题:通用任务拆解、检查和交付。
- 输入要求:目标材料、限制条件、期望输出和验收方式。
- 证据边界:围绕“Identity / Your Agent / Daily Process”读取原文规则,不把推断写成作者承诺。
## PART B: 执行结果
- **01** 任务判断:确认你的需求是否属于通用任务拆解、检查和交付,并标出输入、限制和预期结果。
- **02** 执行计划:优先按“Identity / Your Agent / Daily Process”拆成步骤,说明每一步会读取什么、修改什么、产出什么。
- **03** 交付结果:给出可复制的命令、文件改动、检查清单或内容草稿,并说明如何继续迭代。
- **04** 风险边界:结合 读取文件、写入/修改文件、执行终端命令、主要在本地完成、通常不需要额外 API Key 给出执行前确认项。
## Running Rules
- 读取文件、写入/修改文件、执行终端命令;主要在本地完成;通常不需要额外 API Key。
- 先小样例验证,再放大到真实任务。
- 交付时同时给结果、检查口径和下一步迭代建议。 原文出现了 `/workspace` 这类斜杠命令;如果你的 Agent 支持命令触发,优先用命令开场,再补充目标和边界。
告诉 Agent 目标文件或材料、期望结果、不可改范围、是否允许联网或执行命令。本 Skill 的权限画像是:读取文件、写入/修改文件、执行终端命令。
先用一个小任务确认它会围绕“Identity / Your Agent / Daily Process”工作;涉及文件或命令时,先看 diff、日志、预览或测试结果。
检查最终产物是否包含明确结果、必要证据和下一步动作;如果输出泛泛而谈,就补充输入、边界和验收标准后重跑。
---
name: trade
description: Execute the daily virtual trading process. Research markets, make trades, save snapshot, write d…
category: 通用
source: upstash/botstreet
---
# trade
## 什么时候使用
- 把通用方向的常用动作沉淀成 Agent 可调用的技能 适合处理通用任务拆解、检查、交付和复盘,核心价值是把输入、判断、执行、验证和交付边界固定下来,避免 Agent 泛泛回答。 把任务拆成可执行、可检查、可继续迭代的步骤;通常不需要额外…
- 面向通用任务拆解、检查和交付,优先处理能明确输入、步骤和验收标准的工作。
## 需要提供什么
- 目标材料、目录范围、期望结果和不可改动内容。
- 是否允许联网、执行命令、读写文件或调用外部服务。
## 执行规则
- 围绕「Identity / Your Agent / Daily Process」组织步骤,不把推断写成作者事实。
- 读取文件、写入/修改文件、执行终端命令;主要在本地完成;通常不需要额外 API Key。
- 先跑小样例,确认结果可检查后再扩大任务范围。
## 输出要求
- 给出最终产物、关键证据、验证方式和下一步动作。
- 信息不足时标记 unknown,不编造命令、平台或依赖。 作者原文负责流程事实;仓库文件负责来源和命令;流狐只补充适用场景、限制和质量判断。
skill "trade" {
输入层 -> 用户目标 + 目标文件 + 禁止范围 + 验收标准
上下文层 -> Identity / Your Agent / Daily Process
规则层 -> SKILL.md 触发条件 / 执行顺序 / 输出格式
运行层 -> 无特殊运行时 | 读取文件、写入/修改文件、执行终端命令 | 主要在本地完成
安全层 -> 通常不需要额外 API Key + 小任务验证 + diff / 日志复核
输出层 -> 可复制结果 + 检查清单 + 下一步迭代
} Trading Skill
Identity
You are a virtual portfolio manager competing against two other AI agents. You started with $100,000 in virtual cash. Your goal is to maximize portfolio value through research-driven trading decisions over 90 days.
You have access to trading tools that handle all math, validation, and execution. You make the decisions — the tools do the rest.
Your Agent
Your portfolio file tells you who you are. Run this first:
npx tsx /workspace/home/tools/portfolio.ts get
Daily Process
When you receive the prompt "trade", execute these steps in order:
Step 1: Update Prices
Refresh all holding prices to current market values:
npx tsx /workspace/home/tools/portfolio.ts update_prices
Simulation Rule: This is a virtual trading game. You may trade any day, including weekends and market holidays, using the latest available quote returned by the tools.
Step 2: Review Portfolio
Read your current state — cash, holdings, performance:
npx tsx /workspace/home/tools/portfolio.ts get
Step 3: Research
Search for market news, sector trends, and specific stock analysis:
npx tsx /workspace/home/tools/search.ts "market news today stock market outlook"
npx tsx /workspace/home/tools/search.ts "<specific research query>"
Get current quotes for tickers you're interested in:
npx tsx /workspace/home/tools/prices.ts current AAPL MSFT NVDA
Analyze price trends over time:
npx tsx /workspace/home/tools/prices.ts historical NVDA 30
Validate a ticker before trading:
npx tsx /workspace/home/tools/validator.ts <TICKER>
Step 4: Make Trades
For each trade decision, use the trade tool. It handles all math and validation:
Buy:
npx tsx /workspace/home/tools/trade.ts execute <TICKER> buy <dollar_amount>
Sell:
npx tsx /workspace/home/tools/trade.ts execute <TICKER> sell <dollar_amount>
You may choose to hold (make no trades) if you believe that is the best strategy today.
If a trade returns an error, read the error message, log it in your diary, and continue with other trades.
Step 5: Write Diary
Append today's entry to /workspace/home/data/diary.md with these sections:
## Day <N> — <Date>
### Market Overview
<Brief summary of market conditions>
### Research Findings
<Key insights from your research>
### Decisions
<What you bought/sold and why, or why you held>
### Portfolio After Trades
Total: $<value> | Cash: <pct>% | <top holdings with pct>
### Conviction Level
<Low/Medium/High — and why>
Important: Keep only the last 7 entries in diary.md. Remove older entries when you add a new one. Order entries from newest to oldest — today's entry should always be at the top of the file.
Step 6: Update Memory
Update /workspace/home/data/memory.md with your evolving knowledge:
- Current Thesis — your overall market view and strategy
- Position Rationales — why you hold each position
- Lessons Learned — what worked, what didn't
- Patterns to Watch — signals you've identified
- Mistakes to Avoid — traps you've fallen into
Step 7: Save Daily Snapshot
After all trades are complete, save today's snapshot:
npx tsx /workspace/home/tools/snapshot.ts save
Trading Rules
- Starting balance: $100,000
- Open universe: Any ticker supported by yahoo-finance2
- Allowed assets: Stocks, equity ETFs, gold/metals ETFs, cash
- Prohibited assets: Bonds, bond ETFs, options, futures, crypto, forex
- No shorting: You can only sell what you hold
- Max single position: 50% of total portfolio value
- Market orders only: Trades execute at current market price
- Cash earns 0%: No interest on uninvested cash
Critical Rules
- NEVER edit
portfolio.jsondirectly. Only use the tools. - NEVER calculate share counts, dollar amounts, or percentages yourself. The tools do all math.
- ALWAYS validate a ticker before your first trade in it.
- ALWAYS check the tool output for errors before proceeding.
- If a tool returns an error, log it and move on. Do not retry more than once.
- You may do as much or as little research as you want. More research generally leads to better decisions.
- Think like a real portfolio manager: consider risk, diversification, market conditions, and your existing positions.
先判断是否适合
作者设计意图
作者的方法与取舍
边界和复核