agentleFS
AtmoskunAtmoskun/claude-skills/skills/review/SKILL.md

review

Skill0 starsChanged 9 months agoMITView on GitHub

定期投资回顾 - 聚合 Portfolio、Research、周会、Inbox 活动,生成周/月回顾报告,输出到 Obsidian

192 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use close-spec, design-with-images and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • dzhngSkill

    close-spec

    When a spec under specs/<feature>/ is done shipping (a write-spec build, or any planned task that produced a spec), archive it to specs/done/ and rewrite it from a build-plan into a durable rationale — the why, the principles, the invariants — pointing back to the real code for the how. Use when implementation has landed and the plan no longer matches what shipped, or the user says a feature/spec is finished. Pairs with [write-spec](../write-spec/SKILL.md) (the plan this closes) and [review](../review/SKILL.md) (run before closing).

    Also in 5 projects that use this

  • dzhngSkill

    design-with-images

    Explore UI and visual design with image generation, then iterate the real implementation against the user-selected concept. Use when brainstorming visual directions, offering mockup variants, or implementing an approved generated design; do not stop at the mockups or the first code approximation.

    Also in 5 projects that use this

  • LambenthanSkill

    daily-arxiv

    每日从 arXiv 拉取新论文,过滤相关性,auto-ingest 高优先级论文,检测 SOTA 更新

    Also in 5 projects that use this

  • LambenthanSkill

    exp-eval

    实验判决门:Review LLM 独立评判实验结果 → 4 种判决路径 → 自动更新 claims confidence、ideas status、graph edges

    Also in 5 projects that use this

  • +3 more from dzhng
  • +5 more from Lambenthan

The file

We split this skill into the rules it gives an agent

2 rules. 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 157 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Run the attribution report generator:

    line 309
  2. Distinctive:

    Run the passed tracker:

    line 327

“Common” is measured: kept by at least 6 in 10 of the 157 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

/review - 定期投资回顾

聚合一段时间内的投资活动:交易、研究、周会讨论、收件箱处理,生成结构化回顾报告。

Instructions for Claude

自动创建目录: 如果 ~/Documents/Obsidian Vault/写作/投资回顾/ 不存在,自动创建。

时间范围计算:

  • week = 过去 7 天(从上周一到本周日)
  • month = 过去 30 天(上月同日到今天)
  • YYYY-MM-DD to YYYY-MM-DD = 自定义范围

数据源扫描(全部并行):

  1. Portfolio / Trades
    • 读取 ~/PORTFOLIO/portfolio_monitor/data/trades.json
    • 或查询 portfolio.db(SQLite,用 Python)
    • 提取:新建仓、加仓、减仓、清仓
  2. Research Notes
    • 扫描 研究/研究笔记/ 中日期在范围内的文件
    • 格式: {TICKER}_YYYY-MM-DD.md
  3. Earnings Analysis
    • 扫描 研究/财报分析/ 中日期在范围内的文件
    • 看文件名中的日期
  4. Thesis Updates
    • 检查 ~/PORTFOLIO/portfolio_monitor/research/companies/*/thesis.md
    • 按文件修改时间过滤
  5. 周会 (Weekly Meetings)
    • 扫描 周会/会议实录 YYYY-MM-DD.md 日期在范围内的
    • 读取前 10 行(含会议摘要和提到公司)
  6. 收件箱
    • 统计范围内新增的 inbox 条目
    • 统计 processed: true vs processed: false
    • 提取高频 tickers
  7. Podcast
    • 扫描 信息源/播客/ 中 publish_date 在范围内的
    • 统计已处理 vs 未处理
  8. 13F Institutional Holdings (季度回顾时重点展示)
    • 扫描 ~/Documents/Obsidian Vault/研究/13F 持仓/ 中的分析报告
    • 也检查 ~/13F-CLAUDE/output/*/ 中的 CSV 数据
    • 按持仓 ticker 过滤:哪些机构增持/减持了你持有的股票
    • 格式化为 "Smart Money Activity" 表格
  9. Supply Chain Mentions
    • 扫描 ~/Documents/Obsidian Vault/研究/供应链/ 中的提及报告
    • 也可查询 ~/.claude/skills/supply-chain/data/supply_chain.db: SELECT * FROM mentions WHERE date >= '{start_date}' ORDER BY date
    • 总结:本期新增的供应链提及(谁在财报中提到了什么公司)
  10. ChatGPT Investment Conversations
    • 扫描 ~/Documents/Obsidian Vault/ChatGPT/Investment Research/ 中日期在范围内的文件
    • 提取 ticker 相关的分析讨论
  11. NotebookLM Q&A Activity
    • 读取 ~/.claude/skills/notebooklm/data/history.json
    • 统计范围内的查询次数和涉及的 ticker
    • 总结关键问答(问了什么、得到了什么答案)

When to Use This Skill

  • 用户使用 /review week、/review month 等
  • 用户说"回顾一下这周/这个月的情况"
  • 每周五下午或月末适合运行

Core Workflow

输入时间范围
       ↓
[1] 计算日期范围
    • week/month/custom → start_date, end_date
       ↓
[2] 并行扫描所有数据源
    • Trades, Research, Earnings, Thesis, 周会, 收件箱, Podcast
    • 13F, 供应链, ChatGPT, NotebookLM Q&A
       ↓
[3] 汇总统计
    • 按类别计数
    • 按 ticker 聚合
    • 识别重点关注的公司(多个数据源都提到的)
       ↓
[4] 生成回顾报告
    • 结构化 Markdown
    • 包含 [[wikilinks]] 到相关笔记
       ↓
[5] 生成行动项
    • 未处理的 inbox 项
    • 需要更新的 thesis
    • 需要跟进的周会决策
       ↓
[6] 保存到 Obsidian
    • 路径: ~/Documents/Obsidian Vault/写作/投资回顾/YYYY-MM-DD_{period}_review.md
    • 同时输出摘要到终端

Quick Start

/review week                         # 过去一周回顾
/review month                        # 过去一个月回顾
/review 2026-01-01 to 2026-01-31     # 自定义日期范围
/review quarter                      # 过去一个季度
/review attribution                  # Source attribution report (Phase 4)
/review passed                       # Monthly passed ticker price check (Phase 4)

输出格式

---
created: YYYY-MM-DD
type: review
period: week
start_date: YYYY-MM-DD
end_date: YYYY-MM-DD
tags: [review, weekly]
---

# 投资回顾:YYYY-MM-DD ~ YYYY-MM-DD

> 周度/月度投资活动汇总

## 📊 Portfolio Activity

### 交易记录
| 日期 | 操作 | Ticker | 方向 | 备注 |
|------|------|--------|------|------|
| 2026-02-03 | 新建仓 | NVDA | Long | AI 基础设施 |
| 2026-02-05 | 加仓 | UBER | Long | Q4 超预期 |

### 持仓变化
- **新增:** NVDA
- **加仓:** UBER (+2%)
- **减仓:** 无
- **清仓:** 无

## 📝 研究活动

### 新增研究笔记 (2)
- [[NVDA_2026-02-05]] - 深度研究
- [[UBER_2026-02-04]] - 财报后更新

### 财报分析 (3)
- [[UBER Q4 2025 vs Q3 2025 Claude Analysis]] - UBER 超预期
- [[AAON-US Q4 2025 vs Q3 2025 Analysis]] - 毛利改善
- [[WOSG-GB Q3 2026 vs Q2 2026 Analysis]] - 销售回暖

### Thesis 更新
- `NVDA/thesis.md` - 更新于 2026-02-05(新增 AI agent 叙事)

## 🗓️ 周会要点

### 会议实录 2026-01-03
> AI硬件链条偏多,存储股两周涨40%,美债利率回到高位需警惕
- 关键 tickers: TSM, MU, BIDU, GOOGL, MSFT
- 行动: 跟踪存储/封装链价格验证

## 📬 收件箱统计

| 指标 | 数量 |
|------|------|
| 本期新增 | 12 |
| 已处理 | 5 |
| 未处理 | 7 |
| 高频 Tickers | NVDA(3), TSLA(2) |

## 🎙️ Podcast 活动
- 新增: 5 篇
- 已处理: 2 篇
- 待处理: 3 篇

## 🏦 13F Smart Money Activity (季度回顾)

| Ticker | 机构动向 | 来源 |
|--------|---------|------|
| PM | Einhorn 增持 15% (Q3 '25) | 13F |
| NVDA | 3 家新建仓 | 13F |

## 🔗 Supply Chain Signals

本期新增的供应链提及:
- TSM Q4 财报提到 NVDA CoWoS 产能扩张
- AVGO 提到 AI networking 需求加速

## 💬 ChatGPT & NotebookLM Activity

- ChatGPT 投资对话: 5 篇 (涉及: NVDA, PM, TSM)
- NotebookLM 查询: 8 次 (主要: PM ZYN thesis, NVDA competition)

## 🔑 重点关注公司

多个数据源同时提到的公司:

| Ticker | 出现次数 | 来源 |
|--------|---------|------|
| NVDA | 12 | Research, 周会, Inbox, Podcast, 13F, Supply Chain, ChatGPT |
| UBER | 5 | Earnings, Research, Trade |

## ✅ Next Actions

- [ ] 处理 7 个未读收件箱条目
- [ ] 更新 UBER thesis(Q4 财报后)
- [ ] 处理 3 个未读播客笔记
- [ ] 跟进周会中提到的存储/封装链

Agent Teams Mode (Experimental)

当数据源多、时间范围长(月度/季度)时,可用 Agent Teams 并行化数据采集 + 交叉验证。

启用条件

  • 环境变量 CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1 已设置
  • 用户使用 --team 参数,或 Claude 判断任务复杂度适合(月度/季度回顾自动建议)

团队结构

Lead Agent (Opus) — 协调者:分配任务、去重合并、生成最终报告
  │
  ├── Teammate A: "Portfolio Analyst" (Sonnet)
  │   → 数据源: trades.json, portfolio.db, thesis updates
  │   → 职责: 交易汇总、持仓变化、thesis 更新检测
  │   → 输出: 交易记录表 + 持仓变化列表 + 更新的 thesis 清单
  │
  ├── Teammate B: "Research Scanner" (Sonnet)
  │   → 数据源: 研究笔记, 财报分析, Podcast, 收件箱
  │   → 职责: 研究活动统计、内容摘要、未处理项清单
  │   → 输出: 新增笔记列表 + 财报分析摘要 + Inbox/Podcast 统计
  │
  ├── Teammate C: "External Intelligence" (Sonnet)
  │   → 数据源: 13F holdings, Supply Chain DB, ChatGPT, NotebookLM history
  │   → 职责: 机构动向、供应链信号、AI 对话活动统计
  │   → 输出: Smart Money 表格 + 供应链信号 + ChatGPT/NLM 活动
  │
  └── Teammate D: "Meeting Analyst" (Sonnet)
      → 数据源: 周会/会议实录, 行动项历史
      → 职责: 会议要点提取、行动项跟进状态
      → 输出: 会议摘要 + 未完成行动项

交叉验证(Agent Teams 独有价值)

Teammates 之间直接通信,实现单 agent 模式做不到的交叉验证:

发现方验证方交叉验证内容
A (Portfolio)C (External)A 发现加仓某 ticker → C 检查 13F 是否有机构同步增持/减持
C (External)A (Portfolio)C 发现 13F manager 大幅减持 → A 检查对应 thesis 的 kill criteria
B (Research)D (Meeting)B 发现新 earnings analysis → D 检查周会是否讨论过同一公司
D (Meeting)B (Research)D 提取周会行动项 → B 检查是否有对应 research note 跟进
C (External)B (Research)C 发现供应链新信号 → B 检查是否有相关 podcast 或研究笔记

Lead Agent 额外职责(仅 Team 模式)

  1. 去重 — 多个 teammate 可能报告同一 ticker,Lead 合并为单条
  2. 冲突标记 — 如 Portfolio 显示加仓但 13F 显示机构减持,Lead 标记为 "⚠️ 关注信号"
  3. 重点公司识别 — 跨 teammate 出现 ≥3 次的 ticker 自动升级为"重点关注"
  4. 提前终止 — 如果某 teammate 数据源为空(如本周无周会),Lead 提前终止该 teammate 节省 token

成本对比

模式适用场景预计 Token
单 Agent(默认)周回顾、数据源少~30K
Agent Teams (--team)月度/季度、全数据源~100-150K

建议: Teammates 用 Sonnet(数据拉取),Lead 用 Opus(综合分析)。

Commands Reference

# 单 Agent 模式(默认)
/review week                          # 周回顾
/review month                         # 月回顾
/review quarter                       # 季回顾
/review YYYY-MM-DD to YYYY-MM-DD      # 自定义范围
/review week --focus TICKER           # 聚焦某个 ticker 的周回顾

# Agent Teams 模式
/review week --team                   # Agent Teams 周回顾
/review month --team                  # Agent Teams 月回顾(推荐)
/review quarter --team                # Agent Teams 季回顾(强烈推荐)

# Phase 4 专项
/review attribution                   # Source attribution report
/review passed                        # Monthly passed ticker review

Phase 4: Attribution & Passed Review

/review attribution

Generates a source attribution report showing which information channels produce the best investment ideas.

Workflow:

  1. Run the attribution report generator:
    cd ~/.claude/skills && python -c "
    from shared.attribution_report import generate_attribution_report
    report = generate_attribution_report(save=True)
    print(report)
    "
  2. Report shows: Source → Ideas → Positions → Pass Rate → Avg Return → Win Rate
  3. Includes "Weekly Meeting" as a source channel (from NLM attribution)
  4. Lists unattributed tickers that need idea_source tagging
  5. Saved to Documents/Obsidian Vault/写作/投资回顾/YYYY-MM-DD_attribution_report.md

/review passed

Monthly check on all passed companies + NLM-based discovery of new candidates.

Workflow:

  1. Run the passed tracker:
    cd ~/.claude/skills && python -c "
    from shared.passed_tracker import generate_full_report
    report = generate_full_report(save=True)
    print(report)
    "
  2. Part 1: Price tracking — compares price_at_pass vs. current price for all passed records
  3. Part 2: NLM discovery — queries 投资观点周报 for tickers discussed but not in portfolio/passed
  4. Shows decision accuracy: what % of your passes were "correct" (stock <5% up since pass)
  5. Saved to Documents/Obsidian Vault/写作/投资回顾/YYYY-MM-DD_passed_review.md

与其他 Skills 的关系

Skill关系
/portfolio读取持仓和交易数据
/research统计研究笔记产出
/thesis检查 thesis 更新 + idea_source attribution
/moc回顾中的 ticker 可生成 MOC
/inbox统计 inbox 处理进度
/podcast统计播客处理进度
/13f读取机构持仓变动(季度回顾重点)
/supply-chain读取供应链提及数据
/chatgpt-organizer统计投资相关 ChatGPT 对话
/notebooklm统计 Q&A 查询活动
/flashback回顾中的 ticker 可深入生成 flashback
NotebookLM/review attribution + /review passed 使用 NLM 查询

注意事项

  • trades.json 格式需要先读取确认结构
  • portfolio.db 是 SQLite,可用 Python 查询
  • 周会文件前几行包含结构化摘要,是最重要的提取目标
  • 日期过滤要兼容不同格式(YYYY-MM-DD, created frontmatter, 文件名中的日期)
  • 回顾报告应该以数据驱动,避免主观判断
  • 输出同时到文件和终端(终端版更简洁)

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.