cxcscmu
cxcscmu/SkillLearnBench
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/cxcscmu/SkillLearnBench --skill run2-13f-fund-analysis命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Handles reading, populating, and saving .docx files using the python-docx library. Use this skill for any tasks involving template filling or modifying Word documents.
Perform various data analysis on SEC 13-F and obtain some insights of fund activities such as number of holdings, AUM, and change of holdings between two quarters.
This skill includes search capability in 13F, such as fuzzy search a fund information using possibly inaccurate name, or fuzzy search a stock cusip info using its name.
基于 SOC 职业分类
正在显示 SKILL.md
| name | run2_13f-fund-analysis |
| description | Analyze fund AUM, holdings count, and cross-quarter investment changes from SEC 13F filings. |
Analyze individual fund holdings and compare changes between quarters.
/root/.claude/skills/13f-analyzer/scripts//root/{quarter}/INFOTABLE.tsv, /root/{quarter}/COVERPAGE.tsvpython3 /root/.claude/skills/13f-analyzer/scripts/one_fund_analysis.py \
--accession_number 0001037389-25-000064 \
--quarter 2025-q3
Output fields:
Total number of holdings: all asset classesTotal AUM: in USD dollars (float)Number of stock holdings: equities onlyTotal stock AUM: equity AUM onlypython3 /root/.claude/skills/13f-analyzer/scripts/one_fund_analysis.py \
--quarter 2025-q3 \
--accession_number 0001193125-25-282901 \
--baseline_quarter 2025-q2 \
--baseline_accession_number 0000950123-25-008343
Output: "Top 10 Buys" and "Top 10 Sells" ranked by Abs change (dollar value change).
The script uses hardcoded /root/INFOTABLE.tsv (missing quarter subdirectory). Use this Python workaround instead:
import pandas as pd
quarter = '2025-q3'
cusip = '69608A108'
topk = 3
infotable = pd.read_csv(f'/root/{quarter}/INFOTABLE.tsv', sep='\t', dtype=str)
infotable['VALUE'] = infotable['VALUE'].astype(float)
holding_details = infotable[infotable['CUSIP'] == cusip]
top = (
holding_details.groupby('ACCESSION_NUMBER')
.agg(TOTAL_VALUE=('VALUE', 'sum'))
.sort_values('TOTAL_VALUE', ascending=False)
.head(topk)
)
coverpage = pd.read_csv(f'/root/{quarter}/COVERPAGE.tsv', sep='\t', dtype=str)
for idx, (accession_number, row) in enumerate(top.iterrows()):
fund_info = coverpage[coverpage['ACCESSION_NUMBER'] == accession_number]
fund_name = fund_info['FILINGMANAGER_NAME'].values[0] if not fund_info.empty else 'Unknown'
print(f'Rank {idx+1}: {fund_name} | Value = {row["TOTAL_VALUE"]:.2f}')