13f data analysis

Skill cxcscmu/SkillLearnBench/skills/b1-one-shot-claude-opus-4-6/financial-analysis/13f-data-analysis

[COLM'26] SkillLearnBench is the first benchmark for evaluating continual learning methods that automatically generate agent skills.

Install

npx -y skills add cxcscmu/SkillLearnBench --skill 13f-data-analysis

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

What its author says it does

Copied from the file, not written here

Analyze SEC 13-F filings data including AUM, holdings count, and cross-quarter comparisons using TSV files.

SKILL.md

2.5 KB, as published. Nobody here has run it

SEC 13-F Data Analysis

Data Structure

13-F data is distributed as TSV files:

COVERPAGE.tsv: Filing metadata. Key columns: ACCESSION_NUMBER, FILINGMANAGER_NAME, REPORTCALENDARORQUARTER
INFOTABLE.tsv: Individual holdings. Key columns: ACCESSION_NUMBER, NAMEOFISSUER, TITLEOFCLASS, CUSIP, VALUE (in thousands $), SSHPRNAMT (shares), SSHPRNAMTTYPE
SUMMARYPAGE.tsv: Aggregated data. Key columns: ACCESSION_NUMBER, TABLEENTRYTOTAL (number of holdings), TABLEVALUETOTAL (AUM in thousands $)

Common Operations with Python/Pandas

Load data

import pandas as pd
cover = pd.read_csv('COVERPAGE.tsv', sep='\t')
info = pd.read_csv('INFOTABLE.tsv', sep='\t')
summary = pd.read_csv('SUMMARYPAGE.tsv', sep='\t')

Find fund by name (fuzzy)

from thefuzz import fuzz
cover['score'] = cover['FILINGMANAGER_NAME'].apply(lambda x: fuzz.token_sort_ratio(search_term.lower(), str(x).lower()))
best = cover.loc[cover['score'].idxmax()]
accession = best['ACCESSION_NUMBER']

Get AUM and holdings count

row = summary[summary['ACCESSION_NUMBER'] == accession]
aum = row['TABLEVALUETOTAL'].values[0]  # in thousands
count = row['TABLEENTRYTOTAL'].values[0]

Compare holdings across quarters

q2_holdings = info_q2[info_q2['ACCESSION_NUMBER'] == acc_q2][['CUSIP','VALUE','SSHPRNAMT']]
q3_holdings = info_q3[info_q3['ACCESSION_NUMBER'] == acc_q3][['CUSIP','VALUE','SSHPRNAMT']]
# Group by CUSIP (a fund may hold multiple lots of same stock)
q2_agg = q2_holdings.groupby('CUSIP')['VALUE'].sum().reset_index()
q3_agg = q3_holdings.groupby('CUSIP')['VALUE'].sum().reset_index()
merged = q3_agg.merge(q2_agg, on='CUSIP', how='outer', suffixes=('_q3','_q2')).fillna(0)
merged['change'] = merged['VALUE_q3'] - merged['VALUE_q2']
top_increases = merged.nlargest(5, 'change')

Find all holders of a specific stock (by CUSIP)

holders = info[info['CUSIP'] == target_cusip]
holders_agg = holders.groupby('ACCESSION_NUMBER')['VALUE'].sum().reset_index()
holders_agg = holders_agg.merge(cover[['ACCESSION_NUMBER','FILINGMANAGER_NAME']], on='ACCESSION_NUMBER')
top_holders = holders_agg.nlargest(3, 'VALUE')

Notes

VALUE column is in thousands of dollars
Multiple INFOTABLE rows can exist for same CUSIP within one filing (different share classes, put/call)
SUMMARYPAGE may have fewer rows than COVERPAGE (not all filings have summary)