Back to Details

thesis-tracker

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Codex adapter note

Codex adapter note

This skill is generated from 
skills/thesis-tracker.md
so Claude Code and Codex users share one canonical workflow.
  • Treat 
    $ARGUMENTS
    as the user's request in the current Codex thread.
  • When the source mentions Claude-only surfaces such as Task, Agent, WebSearch, Bash, Read, or Write, use the closest Codex capability available in this session: subagents when available, web search when needed, shell commands for local tools, and normal file edits for workspace files.
  • Use shared project tools from 
    tools/
    in this repository. Prefer running commands from the repository root with paths like 
    python3 tools/financial_rigor.py ...
    ; if the current thread starts outside the repo, locate the actual checkout path first instead of assuming a fixed home-directory path.
  • Before starting research, run the 
    date
    command to confirm today's date; treat it as the baseline for "latest" data and state the data cutoff date in the report header. Never assume the current date from training data.
  • Preserve the research quality rules from 
    AGENTS.md
    : cross-check financial data, use exact arithmetic tools for valuation/math, and clearly label uncertainty and source gaps.
This skill is generated from 
skills/thesis-tracker.md
so Claude Code and Codex users share one canonical workflow.
  • Treat 
    $ARGUMENTS
    as the user's request in the current Codex thread.
  • When the source mentions Claude-only surfaces such as Task, Agent, WebSearch, Bash, Read, or Write, use the closest Codex capability available in this session: subagents when available, web search when needed, shell commands for local tools, and normal file edits for workspace files.
  • Use shared project tools from 
    tools/
    in this repository. Prefer running commands from the repository root with paths like 
    python3 tools/financial_rigor.py ...
    ; if the current thread starts outside the repo, locate the actual checkout path first instead of assuming a fixed home-directory path.
  • Before starting research, run the 
    date
    command to confirm today's date; treat it as the baseline for "latest" data and state the data cutoff date in the report header. Never assume the current date from training data.
  • Preserve the research quality rules from 
    AGENTS.md
    : cross-check financial data, use exact arithmetic tools for valuation/math, and clearly label uncertainty and source gaps.

投资论文追踪：买入后的纪律系统

Investment Thesis Tracker: Post-Purchase Discipline System

对 $ARGUMENTS 执行投资论文追踪检查。
支持输入格式：
  • 公司名
    — 首次使用时建立投资论文，后续使用时追踪检查
  • 公司名 建立论文
    — 强制重新建立投资论文
  • 公司名 季度检查
    — 基于最新财报进行论文检查
"买入只是开始。真正的工作是持有期间的持续跟踪。" —— 李录
"当事实改变时，我就改变想法。你呢？" —— 凯恩斯
Perform investment thesis tracking checks on $ARGUMENTS.
Supported Input Formats:
  • Company Name
    — Establish an investment thesis on first use, perform tracking checks on subsequent uses
  • Company Name Establish Thesis
    — Force re-establishment of the investment thesis
  • Company Name Quarterly Check
    — Perform thesis check based on the latest financial report
"Buying is just the beginning. The real work is continuous tracking during holding." — Li Lu
"When the facts change, I change my mind. What do you do?" — John Maynard Keynes

设计理念

Design Philosophy

大多数投资者的流程是：研究 → 买入 → 祈祷。缺少买入后的系统化跟踪，导致：
  • 该卖的时候舍不得卖（"再等等，会涨回来的"）
  • 不该卖的时候恐慌卖出（"跌了20%，是不是我错了"）
  • 忘记了当初为什么买的（"我买这个是因为什么来着？"）
巴菲特和李录的做法是：买入前就写下卖出条件。然后每个季度检查论文是否完整。
Most investors' process is: Research → Buy → Pray. The lack of systematic tracking after purchase leads to:
  • Reluctance to sell when it's time to ("Wait a bit longer, it will rebound")
  • Panic selling when it shouldn't be done ("It's dropped 20%, was I wrong?")
  • Forgetting why they bought it in the first place ("Why did I buy this again?")
Warren Buffett and Li Lu's approach is: Write down the selling conditions before buying. Then review the thesis every quarter to ensure its integrity.

执行流程

Execution Process

第一步：判断操作模式

Step 1: Determine Operation Mode

检查是否已存在该公司的投资论文文件（
reports/{公司名}-thesis.md
）：
  • 如果不存在 → 进入建立论文模式
  • 如果存在 → 进入追踪检查模式
  • 如果找不到但用户表示已有 → 询问文件路径
Check if the company's investment thesis file (
reports/{Company Name}-thesis.md
) exists:
  • If it doesn't exist → Enter Establish Thesis mode
  • If it exists → Enter Tracking Check mode
  • If not found but the user claims it exists → Ask for the file path

模式A：建立投资论文

Mode A: Establish Investment Thesis

A0：数据收集

A0: Data Collection

使用 WebSearch 获取当前股价、估值指标（PE/PB/股息率）、最新财报核心数据，用于填写估值锚点。如果已有该公司的 
/investment-research
或 
/investment-team
报告，优先从中读取。
使用 
tools/financial_rigor.py verify-valuation
校验估值数据。
Use WebSearch to obtain current stock price, valuation metrics (PE/PB/dividend yield), and core data from the latest financial report for filling in valuation anchors. If there are existing 
/investment-research
or 
/investment-team
reports for the company, prioritize reading from them.
Use 
tools/financial_rigor.py verify-valuation
to validate valuation data.

A1：核心论文（必须用200字以内写清楚）

A1: Core Thesis (Must be written within 200 words)

投资论文必须回答以下5个问题，每个问题一句话：
我以 ___元 买入 ___公司，因为：
1. 这门生意的本质是___，我理解它的赚钱方式
2. 它的护城河是___，而且在变宽/稳定
3. 管理层___，值得信赖的原因是___
4. 当前价格相当于内在价值的___折，安全边际来自___
5. 即使我错了，下行风险可控，因为___
如果5句话写不完整，这个论文本身就有问题——说明买入决策不够清晰。
The investment thesis must answer the following 5 questions, one sentence per question:
I bought ___ company at ___ yuan because:
1. The nature of this business is ___, and I understand how it makes money
2. Its moat is ___, and it is widening/stable
3. The management is ___, and the reason for trust is ___
4. The current price is a ___ discount to intrinsic value, and the margin of safety comes from ___
5. Even if I'm wrong, the downside risk is controllable because ___
If you can't complete these 5 sentences, the thesis itself is problematic — it means the buying decision isn't clear enough.

A2：核心假设清单

A2: Core Assumption List

把投资论文拆解成可验证的具体假设：
#核心假设验证方式验证频率当前状态
1例：收入增速维持15%+季报收入增速每季度🟢 成立
2例：毛利率稳定在60%+季报毛利率每季度🟢 成立
3例：管理层持续回购回购公告/现金流表每季度🟢 成立
4例：竞争对手未取得突破行业数据/竞对财报每半年🟢 成立
5............
通常3-7个假设。太少说明思考不够深入，太多说明论文不够聚焦。
Break down the investment thesis into verifiable specific assumptions:
#Core AssumptionVerification MethodVerification FrequencyCurrent Status
1Example: Revenue growth remains at 15%+Quarterly report revenue growthQuarterly🟢 Valid
2Example: Gross margin stabilizes at 60%+Quarterly report gross marginQuarterly🟢 Valid
3Example: Management continues share repurchasesRepurchase announcements/cash flow statementQuarterly🟢 Valid
4Example: Competitors have not made breakthroughsIndustry data/competitors' financial reportsSemi-annually🟢 Valid
5............
Usually 3-7 assumptions. Too few means insufficient thinking, too many means the thesis is not focused enough.

A3：红线清单（触发任何一条 = 必须重新评估）

A3: Red Line List (Triggering any item = Must re-evaluate)

#红线条件严重程度触发后动作
1例：管理层诚信出问题（财务造假、关联交易）致命立即清仓
2例：核心业务连续2季度收入下滑严重减仓50%，重新评估
3例：护城河被明确突破（竞对获得同等能力）严重启动深度研究，考虑退出
4例：监管政策根本性改变商业模式严重重新评估内在价值
5例：管理层大规模减持（非计划性）警告深入调查原因
段永平："卖出只有三个理由：1.发现买错了；2.公司基本面变了；3.找到了更好的。"
#Red Line ConditionSeverityAction After Trigger
1Example: Management integrity issues (financial fraud, related-party transactions)FatalLiquidate immediately
2Example: Core business revenue declines for 2 consecutive quartersSevereReduce position by 50% and re-evaluate
3Example: Moat is clearly breached (competitors gain equivalent capabilities)SevereLaunch in-depth research and consider exiting
4Example: Regulatory policies fundamentally change the business modelSevereRe-evaluate intrinsic value
5Example: Large-scale management share reduction (unplanned)WarningConduct in-depth investigation into the reason
Duan Yongping: "There are only three reasons to sell: 1. Found out I bought the wrong one; 2. The company's fundamentals have changed; 3. Found a better one."

A4：估值锚点

A4: Valuation Anchors

指标买入时乐观目标中性目标悲观情景
股价
PE
市值
内在价值估算
安全边际
IndicatorAt PurchaseOptimistic TargetNeutral TargetPessimistic Scenario
Stock Price
PE
Market Capitalization
Intrinsic Value Estimate
Margin of Safety

A5：保存论文

A5: Save Thesis

将投资论文写入 
reports/{公司名}-thesis.md
，包含：
  • 建立日期
  • 买入价格和仓位
  • 核心论文（5句话）
  • 核心假设清单
  • 红线清单
  • 估值锚点
  • 追踪记录表（初始为空）
Write the investment thesis to 
reports/{Company Name}-thesis.md
, including:
  • Establishment date
  • Purchase price and position
  • Core thesis (5 sentences)
  • Core assumption list
  • Red line list
  • Valuation anchors
  • Tracking record table (initially empty)

模式B：追踪检查

Mode B: Tracking Check

B1：读取现有论文

B1: Read Existing Thesis

读取 
reports/{公司名}-thesis.md
，加载：
  • 核心论文
  • 核心假设清单
  • 红线清单
  • 上次检查记录
Read 
reports/{Company Name}-thesis.md
and load:
  • Core thesis
  • Core assumption list
  • Red line list
  • Last check record

B2：收集最新数据

B2: Collect Latest Data

使用 WebSearch 收集：
  1. 最新财报数据（如果有新的季报/年报）
  2. 近期重大事件（管理层变动、监管政策、竞争动态）
  3. 当前股价和估值指标
  4. 内部人交易记录（大股东增减持）
Use WebSearch to collect:
  1. Latest financial report data (if new quarterly/annual reports are available)
  2. Recent major events (management changes, regulatory policies, competitive dynamics)
  3. Current stock price and valuation metrics
  4. Insider trading records (major shareholders' share increases/reductions)

B3：逐条检查核心假设

B3: Check Core Assumptions One by One

对每个核心假设，用最新数据验证：
#核心假设上次状态最新证据当前状态变化
1收入增速15%+🟢 成立Q4收入增速12%🟡 边际弱化⚠️
2毛利率60%+🟢 成立毛利率61.2%🟢 成立—
3...............
状态定义：
  • 🟢 成立 — 最新数据支持该假设
  • 🟡 边际弱化 — 数据仍在可接受范围，但趋势不利
  • 🔴 受损 — 数据明确不支持该假设
  • ⚫ 破裂 — 假设已被推翻
Verify each core assumption with the latest data:
#Core AssumptionLast StatusLatest EvidenceCurrent StatusChange
1Revenue growth 15%+🟢 ValidQ4 revenue growth 12%🟡 Marginally Weakened⚠️
2Gross margin 60%+🟢 ValidGross margin 61.2%🟢 Valid—
3...............
Status Definitions:
  • 🟢 Valid — Latest data supports the assumption
  • 🟡 Marginally Weakened — Data is still within acceptable range but the trend is unfavorable
  • 🔴 Impaired — Data clearly does not support the assumption
  • ⚫ Broken — The assumption has been overturned

B4：红线检查

B4: Red Line Check

逐条检查红线清单：
#红线条件是否触发证据
1管理层诚信问题❌ 未触发—
2核心业务连续2季下滑❌ 未触发—
任何一条红线触发 → 在报告中用醒目标注，给出明确的行动建议。
Check each item in the red line list:
#Red Line ConditionTriggered?Evidence
1Management integrity issues❌ Not Triggered—
2Core business revenue declines for 2 consecutive quarters❌ Not Triggered—
If any red line is triggered → Mark prominently in the report and give clear action recommendations.

B5：估值更新

B5: Valuation Update

指标买入时上次检查当前变化
股价
PE(TTM)
内在价值估算
安全边际
IndicatorAt PurchaseLast CheckCurrentChange
Stock Price
PE(TTM)
Intrinsic Value Estimate
Margin of Safety

B6：输出追踪报告

B6: Output Tracking Report

报告结构

Report Structure

一、论文健康度评分（满分10分）
二、核心假设检查结果（表格）
三、红线检查结果（表格）
四、本期关键变化（不超过500字）
五、估值更新
六、结论与行动建议
七、下次检查需关注的重点
1. Thesis Health Score (Full Score 10)
2. Core Assumption Check Results (Table)
3. Red Line Check Results (Table)
4. Key Changes This Period (No more than 500 words)
5. Valuation Update
6. Conclusion and Action Recommendations
7. Key Focus Areas for Next Check

论文健康度评分标准

Thesis Health Score Criteria

评分含义建议动作
计算公式：健康度 = 10 - (⚫破裂假设数×3) - (🔴受损假设数×2) - (🟡弱化假设数×1) - (红线触发数×5)，最低1分最高10分。
评分含义建议动作
9-10所有假设成立，论文比买入时更强考虑加仓
7-8核心假设成立，个别边际弱化继续持有
5-61-2个假设受损，但核心逻辑未变持有但提高警惕
3-4多个假设受损，论文基础动摇考虑减仓
1-2红线触发或核心假设破裂强烈建议卖出
Calculation Formula: Health Score = 10 - (Number of ⚫ Broken Assumptions ×3) - (Number of 🔴 Impaired Assumptions ×2) - (Number of 🟡 Weakened Assumptions ×1) - (Number of Triggered Red Lines ×5), minimum 1 point, maximum 10 points.
ScoreMeaningRecommended Action
9-10All assumptions are valid, thesis is stronger than at purchaseConsider increasing position
7-8Core assumptions are valid,个别 marginally weakenedContinue holding
5-61-2 assumptions are impaired, but core logic remains unchangedHold but increase vigilance
3-4Multiple assumptions are impaired, thesis foundation is shakenConsider reducing position
1-2Red line triggered or core assumptions brokenStrongly recommend selling

结论必须明确回答

Conclusion Must Clearly Answer

  1. 论文还完整吗？ 完整 / 边际弱化 / 受损 / 破裂
  2. 该怎么做？ 加仓 / 持有 / 减仓 / 清仓
  3. 下次检查时间：下一个季报发布后 / 某个特定事件后
  1. Is the thesis still intact? Intact / Marginally Weakened / Impaired / Broken
  2. What should be done? Increase Position / Hold / Reduce Position / Liquidate
  3. Next Check Time: After next quarterly report release / After a specific event

B7：更新论文文件

B7: Update Thesis File

将本次检查记录追加到 
reports/{公司名}-thesis.md
的追踪记录表中：
检查日期健康度核心变化动作建议
2026-04-097/10收入增速放缓至12%，但利润率改善持有
Append this check record to the tracking record table in 
reports/{Company Name}-thesis.md
:
Check DateHealth ScoreCore ChangesAction Recommendation
2026-04-097/10Revenue growth slowed to 12%, but profit margin improvedHold

关键原则

Key Principles

  • 买入前就写好卖出条件 — 冷静时做的决策比恐慌时做的好
  • 论文要具体到可验证 — "公司很好"不是论文，"ROE>25%且趋势稳定"才是
  • 红线一旦触发就行动 — 最怕的是"再等等看"，这是亏大钱的开始
  • 论文破裂 ≠ 股价下跌 — 股价跌30%不一定要卖，论文破裂才要卖
  • 诚实面对错误 — 论文建错了就承认，不要为了面子硬撑
  • Write selling conditions before buying — Decisions made calmly are better than those made in panic
  • Thesis must be specific enough to be verifiable — "The company is good" is not a thesis; "ROE>25% and trend is stable" is
  • Act immediately when red line is triggered — The worst thing is "Wait and see", which is the start of big losses
  • Thesis broken ≠ Stock price drop — You don't have to sell if the stock drops 30%, only when the thesis is broken
  • Be honest about mistakes — Admit if the thesis was built wrong, don't hold on for face