sec-edgar-parser

name	sec-edgar-parser
description	Skill for accessing, downloading, and parsing financial filings from the SEC EDGAR database. Helps users retrieve 10-K, 10-Q, 8-K, and other forms using the SEC's JSON API.

SEC EDGAR Parser

This skill enables the agent to interact with the SEC's EDGAR system to retrieve corporate filings and extract financial data.

Workflow

1. Identify Company & Filing

• Find CIK: Use the company ticker (e.g., AAPL, TSLA) to find the Central Index Key (CIK). The SEC maps tickers to CIKs via a public JSON file (company_tickers.json).
• Select Filing Type: Determine which form is needed (e.g., 10-K for annual reports, 10-Q for quarterly, 8-K for current events).

2. Fetch Filing Metadata

• Use the SEC's submission history API: https://data.sec.gov/submissions/CIK{cik_padded}.json
• This JSON contains a list of all recent filings with their accession numbers, filing dates, and form types.

3. Locate & Download Document

• Construct the URL for the specific filing document. The format typically involves the CIK (without leading zeros) and the Accession Number (with hyphens removed).
• HTML URL Pattern: https://www.sec.gov/Archives/edgar/data/{cik}/{accession_number}/{primary_document}
• Interactive Data (XBRL): Often available as XML or JSON.

4. Parse Content

• Text Extraction: Use BeautifulSoup to strip HTML tags and extract sections (e.g., "Item 1A. Risk Factors", "Item 7. MD&A").
• Table Extraction: Use pd.read_html() to convert HTML tables into Pandas DataFrames.

Key Requirements

• User-Agent: The SEC requires a valid User-Agent header in the format Sample Company Name [e-mail]. Requests without this will be blocked.
• Rate Limiting: The SEC limits requests to ~10 per second. Scripts should include slight delays if making many requests.

Bundled Scripts

scripts/sec_utils.py

A utility script to:

1. Get CIK from Ticker.
2. Get recent filings list for a company.
3. Download a specific filing's text.

Example Usage

Get Recent 10-K Filings

from scripts.sec_utils import get_cik, get_filings

cik = get_cik("AAPL")
filings = get_filings(cik, form_type="10-K")
print(filings.head())

Download a Filing

from scripts.sec_utils import download_filing
# Use the URL from the filings dataframe
text = download_filing(filings.iloc[0]['url'])