kanchi-dividend-sop by tradermonty/claude-trading-skills
npx skills add https://github.com/tradermonty/claude-trading-skills --skill kanchi-dividend-sop将 Kanchi 的五步法实现为美国股息投资的确定性工作流程。优先考虑安全性和可重复性,而非追逐激进收益率。
当用户需要以下内容时,使用此技能:
入场信号脚本需要 FMP API 访问权限:
export FMP_API_KEY=your_api_key_here
在运行工作流程前,准备以下输入之一:
skills/value-dividend-screener/scripts/screen_dividend_stocks.py 的输出。skills/dividend-growth-pullback-screener/scripts/screen_dividend_growth.py 的输出。使用 --input 时,请提供以下格式之一的 JSON:
广告位招租
在这里展示您的产品或服务
触达数万 AI 开发者,精准高效
{
"profile": "balanced",
"candidates": [
{"ticker": "JNJ", "bucket": "core"},
{"ticker": "O", "bucket": "satellite"}
]
}
或简化格式:
{
"tickers": ["JNJ", "PG", "KO"]
}
为了生成确定性产物,请提供股票代码给:
python3 skills/kanchi-dividend-sop/scripts/build_sop_plan.py \
--tickers "JNJ,PG,KO" \
--output-dir reports/
对于第 5 步入场时机产物:
python3 skills/kanchi-dividend-sop/scripts/build_entry_signals.py \
--tickers "JNJ,PG,KO" \
--alpha-pp 0.5 \
--output-dir reports/
首先收集并锁定参数:
加载 references/default-thresholds.md 并应用基准设置,除非用户覆盖。
从一个偏向质量的股票池开始:
使用明确的来源优先级收集股票代码:
skills/value-dividend-screener/scripts/screen_dividend_stocks.py 输出(FMP/FINVIZ)。skills/dividend-growth-pullback-screener/scripts/screen_dividend_growth_rsi.py 输出。在继续之前,返回按组分组的股票代码列表。
主要规则:
forward_dividend_yield >= 3.5%陷阱控制:
>= 8%)标记为 deep-dive-required。输出:
PASS 或 FAIL。deep-dive-required 标志。要求:
添加安全检查:
当趋势混杂但未破坏时,归类为 HOLD-FOR-REVIEW 而非硬性拒绝。
使用 references/valuation-and-one-off-checks.md 并应用特定板块的估值逻辑:
PER x PBR 可保持为主要方法。P/FFO 或 P/AFFO 代替普通的 P/E。P/E、P/FCF 和历史范围。始终报告每个股票代码使用了哪种估值方法。
拒绝或降级那些近期利润依赖一次性效应的股票:
为每个 FAIL 记录一行证据以保持可审计性。
机械地设置入场触发条件:
+0.5pp)。P/E、P/FFO 或 P/FCF)。执行模式:
40% -> 30% -> 30%。始终生成三个产物:
PASS、HOLD-FOR-REVIEW、FAIL 及证据)。references/stock-note-template.md)。返回和/或生成:
references/stock-note-template.md 的承销备忘录集。skills/kanchi-dividend-sop/scripts/build_sop_plan.py 在 reports/ 中生成的可选计划产物文件。skills/kanchi-dividend-sop/scripts/build_entry_signals.py 在 reports/ 中生成的可选第 5 步入场信号产物。使用此最低节奏:
首先运行此技能,然后将输出交接给:
kanchi-dividend-review-monitor 用于每日/每周/每季度的异常检测。kanchi-dividend-us-tax-accounting 用于账户定位和税务分类规划。skills/kanchi-dividend-sop/scripts/build_sop_plan.py:确定性 SOP 计划生成器。skills/kanchi-dividend-sop/scripts/tests/test_build_sop_plan.py:计划生成测试。skills/kanchi-dividend-sop/scripts/build_entry_signals.py:第 5 步目标买入计算器(5y avg yield + alpha)。skills/kanchi-dividend-sop/scripts/tests/test_build_entry_signals.py:信号计算测试。references/default-thresholds.md:基准阈值和配置文件调优。references/valuation-and-one-off-checks.md:板块估值映射和一次性事件检查清单。references/stock-note-template.md:每个候选股票的一页式备忘录模板。每周安装次数
76
代码仓库
GitHub 星标数
394
首次出现
2026年2月23日
安全审计
安装于
cursor73
github-copilot72
codex72
amp72
kimi-cli72
gemini-cli72
Implement Kanchi's 5-step method as a deterministic workflow for US dividend investing. Prioritize safety and repeatability over aggressive yield chasing.
Use this skill when the user needs:
The entry signal script requires FMP API access:
export FMP_API_KEY=your_api_key_here
Prepare one of the following inputs before running the workflow:
skills/value-dividend-screener/scripts/screen_dividend_stocks.py.skills/dividend-growth-pullback-screener/scripts/screen_dividend_growth.py.When using --input, provide JSON in one of these formats:
{
"profile": "balanced",
"candidates": [
{"ticker": "JNJ", "bucket": "core"},
{"ticker": "O", "bucket": "satellite"}
]
}
Or simplified:
{
"tickers": ["JNJ", "PG", "KO"]
}
For deterministic artifact generation, provide tickers to:
python3 skills/kanchi-dividend-sop/scripts/build_sop_plan.py \
--tickers "JNJ,PG,KO" \
--output-dir reports/
For Step 5 entry timing artifacts:
python3 skills/kanchi-dividend-sop/scripts/build_entry_signals.py \
--tickers "JNJ,PG,KO" \
--alpha-pp 0.5 \
--output-dir reports/
Collect and lock the parameters first:
Load references/default-thresholds.md and apply baseline settings unless the user overrides.
Start with a quality-biased universe:
Use explicit source priority for ticker collection:
skills/value-dividend-screener/scripts/screen_dividend_stocks.py output (FMP/FINVIZ).skills/dividend-growth-pullback-screener/scripts/screen_dividend_growth_rsi.py output.Return a ticker list grouped by bucket before moving forward.
Primary rule:
forward_dividend_yield >= 3.5%Trap controls:
>= 8%) as deep-dive-required.Output:
PASS or FAIL per ticker.deep-dive-required flag for potential yield traps.Require:
Add safety checks:
When trend is mixed but not broken, classify as HOLD-FOR-REVIEW instead of hard reject.
Use references/valuation-and-one-off-checks.md and apply sector-specific valuation logic:
PER x PBR can remain primary.P/FFO or P/AFFO instead of plain P/E.P/E, P/FCF, and historical range.Always report which valuation method was used for each ticker.
Reject or downgrade names where recent profits rely on one-time effects:
Record one-line evidence for each FAIL to keep auditability.
Set entry triggers mechanically:
+0.5pp).P/E, P/FFO, or P/FCF).Execution pattern:
40% -> 30% -> 30%.Always produce three artifacts:
PASS, HOLD-FOR-REVIEW, FAIL with evidence).references/stock-note-template.md).Return and/or generate:
references/stock-note-template.md.skills/kanchi-dividend-sop/scripts/build_sop_plan.py in reports/.skills/kanchi-dividend-sop/scripts/build_entry_signals.py in reports/.Use this minimum rhythm:
Run this skill first, then hand off outputs:
kanchi-dividend-review-monitor for daily/weekly/quarterly anomaly detection.kanchi-dividend-us-tax-accounting for account-location and tax classification planning.skills/kanchi-dividend-sop/scripts/build_sop_plan.py: deterministic SOP plan generator.skills/kanchi-dividend-sop/scripts/tests/test_build_sop_plan.py: tests for plan generation.skills/kanchi-dividend-sop/scripts/build_entry_signals.py: Step 5 target-buy calculator (5y avg yield + alpha).skills/kanchi-dividend-sop/scripts/tests/test_build_entry_signals.py: tests for signal calculations.references/default-thresholds.md: baseline thresholds and profile tuning.references/valuation-and-one-off-checks.md: sector valuation map and one-off checklist.references/stock-note-template.md: one-page memo template for each candidate.Weekly Installs
76
Repository
GitHub Stars
394
First Seen
Feb 23, 2026
Security Audits
Gen Agent Trust HubPassSocketPassSnykWarn
Installed on
cursor73
github-copilot72
codex72
amp72
kimi-cli72
gemini-cli72
Python PDF处理教程:合并拆分、提取文本表格、创建PDF文件
65,000 周安装