Run3 sec 13f search fund coverpage
How to fuzzy search for a fund in COVERPAGE.tsv to find its accession_number, using either built-in scripts or manual searchFrom its SKILL.md
npx -y skills add cxcscmu/SkillLearnBench --skill run3_sec-13f-search-fund-coverpageAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
SKILL.md
1.7 KB, 411 tokens by cl100k_base, as published. Nobody here has run it
Searching for a Fund in COVERPAGE
Method 1: Use Built-in Scripts (Preferred)
# Check if search script exists
cat /root/2025-q3/scripts/search_fund.py 2>/dev/null
# Run it if available
python /root/2025-q3/scripts/search_fund.py "renaissance technologies" 2>/dev/null
python /root/2025-q3/scripts/search_fund.py "berkshire hathaway" 2>/dev/null
Method 2: Manual Fuzzy Search in COVERPAGE.tsv
import pandas as pd
cover = pd.read_csv('/root/2025-q3/COVERPAGE.tsv', sep='\t', low_memory=False)
# Print columns to understand structure
print(cover.columns.tolist())
# Fuzzy search by fund name - check various possible column names
# Common column names: FILINGMANAGER_NAME, COMPANY_NAME, etc.
search_term = "renaissance technologies"
for col in cover.columns:
if 'name' in col.lower() or 'manager' in col.lower() or 'company' in col.lower():
print(f"Column: {col}")
matches = cover[cover[col].astype(str).str.lower().str.contains(search_term, na=False)]
if len(matches) > 0:
print(matches[[col, 'ACCESSION_NUMBER']].head(10))
Getting the Accession Number
The accession_number from COVERPAGE links to INFOTABLE for holdings data.
# Example: find Renaissance Technologies
search = "renaissance"
name_col = 'FILINGMANAGER_NAME' # adjust based on actual column name
matches = cover[cover[name_col].astype(str).str.lower().str.contains(search, na=False)]
print(matches[['ACCESSION_NUMBER', name_col]].drop_duplicates())
# Pick the best matching accession_number
accession = matches['ACCESSION_NUMBER'].iloc[0]
What ships with it
Read from the repository
Just SKILL.md. No reference files, no scripts.