Bankruptcy Court Monitoring Toolkit

Open-source Python tools for monitoring US bankruptcy courts and detecting patterns in publicly available PACER data. Standard library only. No pip install needed.

Three tool families:

1. RSS Monitor -- Real-time court surveillance using free RSS feeds ($0/day)
2. Discharge Bar Screener -- Detect potential 11 U.S.C. 1328(f) violations, 109(g) filing bars, and Ch.7-to-Ch.13 pipeline patterns in PACER CSV exports
3. Practice Report -- Attorney practice quality profiles: volume, outcomes, case duration, statutory compliance, repeat debtors, comparative analysis

RSS Monitor

rss_monitor.py -- Real-Time Court Surveillance

Fetches free public RSS feeds from US bankruptcy courts and runs five analysis layers on every new docket entry. No PACER login required. $0 operating cost. Standard library only.

python rss_monitor.py                         # Default courts from config
python rss_monitor.py --court mowbk           # Single court
python rss_monitor.py --courts mowbk ksbk     # Multiple courts
python rss_monitor.py --national              # National mill-detection mode
python rss_monitor.py --fullsend              # ALL ~90 US bankruptcy courts
python rss_monitor.py --screen                # Enable 1328(f) screening
python rss_monitor.py --expanded              # Expanded court set from config
python rss_monitor.py --portfolio             # Show all portfolio hits
python rss_monitor.py --full                  # All display flags on
python rss_monitor.py --all                   # Show every RSS entry
python rss_monitor.py --stats                 # Portfolio loading stats
python rss_monitor.py --rebuild               # Force rebuild portfolio cache
python rss_monitor.py --dry-run               # Fetch without updating state
python rss_monitor.py --reset                 # Clear seen-state

Five surveillance layers:

1. Case-specific alerts -- Hand-picked cases with custom trigger keywords (e.g., "motion to dismiss", "withdrawal", "contempt"). Configurable per case.
2. Portfolio monitoring -- Cross-references RSS entries against PACER CSV exports. Any docket activity on a target attorney's case triggers an alert.
3. New filing detection -- Catches voluntary petitions as they hit the docket. Track new filings in real time.
4. Section 1328(f) screening -- Screens new Ch.13 filings against the debtor index to flag discharge-barred cases the moment they appear on the docket.
5. Serial filer detection -- Matches debtor names across the portfolio to identify repeat filers.

Additional features:

• Mill signal scoring (heuristic scoring of docket events for bankruptcy mill indicators)
• Tiered priority system (T1-T6: primary, secondary, watchlist, portfolio, associated, control)
• Filing velocity trends and time-series tracking (--timeseries)
• Monitored case changelog (--history)
• Covers all ~90 US bankruptcy courts

Configuration

Copy rss_config.example.json to rss_config.json and customize:

{
  "courts": {
    "default": ["mowbk", "ksbk"],
    "expanded": ["mowbk", "ksbk", "moebk"],
    "national": ["mowbk", "ksbk", "txsb", "ilnb", "cacb", "flsb"]
  },
  "target_attorneys": {
    "names": ["Smith", "Jones"]
  },
  "watched_cases": {
    "24-12345": {
      "name": "Example Debtor",
      "court": "mowbk",
      "tier": 1,
      "attorney": "Smith, John",
      "triggers": ["withdrawal", "plan modification", "contempt"]
    }
  },
  "portfolio_triggers": {
    "keywords": ["motion to dismiss", "conversion", "sanctions", "withdrawal"]
  }
}

See rss_config.example.json for the full configuration template with all options.

How RSS Feeds Work

Every US bankruptcy court publishes a free RSS feed at a predictable URL:

https://ecf.{court-code}.uscourts.gov/cgi-bin/rss_outside.pl

These feeds contain the last ~24 hours of docket activity. Entries disappear when they roll off, so the monitor saves state between runs to track what's been seen. No authentication needed. No PACER charges.

Data Directory

The monitor stores state in rss_data/ (auto-created):

• rss_state.json -- Seen entries (prevents duplicate alerts)
• rss_changelog.json -- History of monitored case activity
• rss_alerts.json -- Triggered alerts log
• portfolio_cache.json -- Cached portfolio from CSV exports
• timeseries.json -- Filing velocity data

---

Discharge Bar Screener

What is Section 1328(f)?

Section 1328(f) of the Bankruptcy Code bars discharge in a Chapter 13 case if the debtor received a prior discharge within certain time limits:

• 1328(f)(1): No discharge if a Ch.7, Ch.11, or Ch.12 discharge was granted within 4 years before the Ch.13 filing date.
• 1328(f)(2): No discharge if a Ch.13 discharge was granted within 2 years before the Ch.13 filing date.

This provision was enacted by BAPCPA (effective October 17, 2005).

Why this matters: When an attorney files a Ch.13 case that is discharge-barred, the debtor pays filing fees and attorney fees for a case that can never achieve its purpose. The debtor's time is wasted, their filing history is impacted, and they receive no benefit. This is a bright-line test -- three data points determine the violation:

1. Date of prior discharge
2. Date of new Ch.13 filing
3. Chapter of prior case

No expert testimony needed. No subjective judgment. Arithmetic.

screen_1328f.py -- Core Screener

The primary tool. Screens PACER Case Locator CSV exports for 1328(f) violations.

python screen_1328f.py --data-dir ./csv-exports --target Smith_John
python screen_1328f.py --data-dir ./csv-exports --target Smith_John --target Doe_Jane
python screen_1328f.py --data-dir ./csv-exports --target Smith_John --control Jones_Bob
python screen_1328f.py --data-dir ./csv-exports --target Smith_John --output-json results.json

What it does:

• Loads all PACER CSV exports from the specified directory
• Groups cases by debtor name (fuzzy matching on first + last name)
• For each debtor with 2+ cases, checks if any Ch.13 filing falls within the 1328(f) statutory window after a prior discharge
• Reports violations with full case details, gap calculations, and attorney attribution
• Optionally outputs structured JSON for further analysis

Output sections:

• f(1) violations (4-year bar: Ch.7/11/12 discharge -> Ch.13)
• f(2) violations (2-year bar: Ch.13 discharge -> Ch.13)
• Cross-attorney detail (prior case by different attorney)
• Most egregious cases (shortest gaps, non-dismissed barred cases)
• Per-attorney summary table
• Methodological caveats

screen_discharge_bars.py -- Multi-Statute Scanner

Extended analysis covering three related statutory bars:

python screen_discharge_bars.py --data-dir ./csv-exports --target Smith_John

Three analyses:

1. Section 1328(f) refinements -- Same-firm vs. cross-firm classification, time-gap histograms, strict name matching for "discharged despite bar" cases
2. Section 109(g) filing bar -- Cases dismissed then refiled within 180 days by the same attorney
3. Ch.7 -> Ch.13 pipeline -- Same attorney handles Ch.7 discharge then files Ch.13 for same debtor (business model indicator)

---

Attorney Practice Report

practice_report.py -- Practice Quality Profiles

Same PACER CSVs, same --target / --control pattern. Generates a comprehensive practice profile for any attorney in your data.

python practice_report.py --data-dir ./csv-exports --target Smith_John
python practice_report.py --data-dir ./csv-exports --target Smith_John --control Jones_Bob
python practice_report.py --data-dir ./csv-exports --target Smith_John --output-json report.json
python practice_report.py --data-dir ./csv-exports --target Smith_John --oneline
python practice_report.py --data-dir ./csv-exports --all

Six metric groups:

1. Volume Profile -- Total cases, chapter breakdown (7/11/12/13), chapter mix %, filing date range, filing velocity (cases/year), courts active
2. Outcome Analysis -- Ch. 13 dismissal/discharge/pending rates, disposition breakdown (top reasons: failure to make payments, failure to file, etc.)
3. Case Duration -- Median days to dismissal, median days to discharge, % dismissed within 90 days (bare petition signal), % dismissed within 180 days
4. Statutory Compliance -- 1328(f)(1) and (2) hits with rate per 1,000 Ch. 13 cases, 109(g) refiling bar hits, Ch. 7 → Ch. 13 pipeline count
5. Debtor Patterns -- Unique debtors, debtors with 2+ cases under same attorney, serial filing rate, max cases per debtor
6. Comparative Analysis -- Side-by-side table with delta column when comparing target vs control

Output formats:

• Text (default): Clean tables, one section per metric group
• JSON (--output-json): Structured for aggregation across districts
• One-line (--oneline): Quick scan, one line per attorney: Smith, John | 450 Ch.13 | 78% dismissed | 2.3/yr | 12 1328f hits (2.7%)

---

Getting PACER Data

These tools process locally-downloaded CSV files. They do not access PACER servers or make any network calls.

See docs/pacer_csv_guide.md for step-by-step instructions on downloading CSVs from the PACER Case Locator.

Quick version:

1. Go to https://pcl.uscourts.gov
2. Create a free PACER account (if you don't have one)
3. Search by attorney name, selecting bankruptcy courts
4. Export results as CSV
5. Save CSVs to a directory
6. Run the screener against that directory

CSV files should follow the naming convention: api_{LastName_FirstName}_{court}_{timestamp}.csv

Quick Start

git clone https://github.com/ilikemath9999/bankruptcy-discharge-screener.git
cd bankruptcy-discharge-screener

# RSS Monitor
cp rss_config.example.json rss_config.json   # Edit with your courts/attorneys
python rss_monitor.py --help

# Discharge Bar Screener
python screen_1328f.py --data-dir ./csv-exports --help

Requirements

• Python 3.8 or higher
• Standard library only -- no pip install needed

Processes 100K+ cases in under a minute on commodity hardware.

Methodology

Name Matching

Debtor names are normalized (lowercased, suffixes stripped, middle names handled) and matched on first + last name. This catches middle-name variations and suffix differences but may produce false positives for common names. All hits require manual verification.

BAPCPA Filter

Only Ch.13 cases filed on or after October 17, 2005 are screened. Section 1328(f) did not exist before BAPCPA.

Deduplication

Cases are deduplicated by PACER case ID. Joint cases ("John Smith and Jane Smith") are split and each spouse is matched independently.

Known Limitations

• Fuzzy name matching can produce false positives (different people, same name)
• PACER data may have entry lag or date imprecision
• "Discharged despite bar" cases may have legitimate explanations (hardship discharge, court waiver)
• The bar prevents discharge, not filing -- an open case is not necessarily an error yet

Live Map

View the national map -- see which districts have been screened and what the numbers look like.

The map updates as volunteers screen new districts. Every row is independently generated from public PACER data.

---

Screen Your District

We tested this on two federal districts, 56,563 Ch. 13 filings screened, 360 flagged inside the 1328(f) window, 14 hand-verified, all confirmed. 114 flagged cases in the multi-district sample received a discharge despite the bar.

Help build the national picture:

• Screening guide, step-by-step: pull PACER CSVs, run the screener, share results (~30 min per district)
• Submit results, form field reference for reporting what you find
• Leaderboard, districts screened, records, hall of fame

Pick a district, pick some high-volume filers, and run the tool. The more districts screened, the clearer the picture of whether this is localized or systemic. First person to find a district with a 5%+ hit rate gets the top spot.

---

Running Tests

python -m unittest discover tests

License

MIT License. See LICENSE.

Disclaimer

This tool is for research and analysis purposes. It identifies potential violations based on publicly available data. All results require manual verification before drawing conclusions. The tool does not provide legal advice.