README.md

---
pretty_name: "Breach Clock (California data-breach notice index)"
license: other
license_name: mit-code-cc0-1.0-data
license_link: https://huggingface.co/datasets/CyberNative-AI/breach-clock/blob/main/DATASET.md
language:
  - en
tags:
  - data-breach
  - california
  - public-records
  - legal
  - index
size_categories:
  - n<1K
---

# Breach Clock

Breach Clock builds a small, static index of protection offers and printed enrolment deadlines from California Attorney General data-breach notice filings. It is a search aid, not an eligibility decision: a person’s own letter and its activation code determine whether they can enrol.

A browsable version of this index: https://cybernative.ai/products/breach-clock/

## Run

Requires Python 3.11+ and `pdftotext` from Poppler. The scripts use only the Python standard library and make no metered API calls.

```sh
python scripts/fetch_registry.py
python scripts/fetch_notices.py --only-id sb24-626307 --max-reports 1
python scripts/extract_offers.py
python scripts/build_index.py
python -m unittest discover -s tests -v
```

The downloaded source inputs and `build/source-manifest.json` are local build inputs; they are intentionally not committed. Notice PDFs and notice prose are never published by this repository.

## Output

`data/offers.jsonl` is the line-oriented index. `data/offers.json` is the browser-ready equivalent. `data/coverage.json` describes the exact build scope and counts. Each row links to the original report and notice, and records the notice hash and access time.

`deadline_status` was set once, at the build time recorded in `data/coverage.json` (2026-07-30). It is not recomputed. Before relying on a row marked `open`, compare its `enrollment_deadline` with today's date. On 2026-09-27, 8 of the 20 rows marked `open` had a deadline that had already passed; the latest deadline among them is 2026-10-31. To count them for today's date:

```sh
python -c "import json,datetime; t=str(datetime.date.today()); rows=[json.loads(l) for l in open('data/offers.jsonl')]; print(sum(r['deadline_status']=='open' and r['enrollment_deadline']<t for r in rows), 'of', sum(r['deadline_status']=='open' for r in rows))"
```

The software is MIT-licensed. CyberNative AI LLC dedicates only its rights, if any, in the selection and arrangement of `data/` under CC0 1.0 Universal; linked or underlying notices, notice prose, third-party trademarks, and other third-party material are excluded. [DATASET.md](DATASET.md) governs the exact scope.