What shipped
Built a Python 3.11+ standard-library CLI package named coordinated_spend_watchboard. It creates a deterministic demo workspace, generates campaign-finance watchboard artifacts, validates required outputs and safety boundaries, and includes a small local delta comparison command.
Architecture
- Reused the parent build's stable local-workspace pattern: sample inputs, generated artifacts, machine-readable JSON, Markdown briefings, CSV scorecard, and validator-enforced notices.
- Kept the implementation standard-library only with
argparse,csv,json,datetime,pathlib, andunittest. - Used fixed generation dates so scoring and tests remain deterministic.
- Made validation part of the product boundary by rejecting missing notices and prohibited legal-conclusion phrasing.
Trimmed scope
- No hosted dashboard, database, live FEC calls, state portal calls, scraping, OAuth, API keys, browser automation, or provider/model calls.
- The scoring model is intentionally explainable and simple rather than predictive.
- The sample data is synthetic and intended for workflow demonstration, not factual race monitoring.
Limitations
- Real-world field normalization across FEC and state exports will require adapters per jurisdiction.
- The validator checks artifact structure and safety language, but it cannot verify whether source URLs are current or complete.
- The pressure score prioritizes review attention; it is not a compliance, legal, or factual finding.
Suggested next steps
- Add importer templates for common FEC CSV exports and one state disclosure export.
- Add a source-receipt folder convention for preserving downloaded originals.
- Add before/after snapshot fixtures to make the delta workflow more useful for weekly refreshes.
- Add optional jurisdiction-specific filing-calendar templates maintained as local files.