Finance Labs v0.1.0 — ten offline diagnostics

First versioned release of ten independent offline diagnostics, with Python wheels, source distributions, and reproducible example output.

Try one lab

git clone --branch v0.1.0 --depth 1 https://github.com/Jo2234/finance-labs.git
cd finance-labs
python3 -m venv .venv
source .venv/bin/activate
python -m pip install ./labs/margin-cascade
margin-cascade --book labs/margin-cascade/examples/sample_book.csv --shock ALPHA=-0.22 --shock GAMMA=-0.08 --format markdown

Requires Python 3.10+. The root is a collection, not an installable package. To install a downloaded wheel offline: python -m pip install --no-index path/to/package.whl. The source archives and repository include each lab's synthetic sample inputs.

Included tools

Prompt Firewall, ETF Liquidity, Margin Cascade, Factor Crowding, Bond Auction Tail, Options Skew, Covenant Headroom, Guidance Drift, Microstructure Stress, and Central Bank Path. Every lab has empty runtime dependencies and its own CLI and README.

Validation and limits

All 130 local tests passed. Ten source distributions built, ten wheels built from those source distributions, license/sample inclusion checks passed, and every documented CLI example ran in a fresh wheel-only environment. validation.json contains the actual outputs and SHA256SUMS covers every uploaded distribution and validation file. CI repeats packaging and installed examples and runs all lab tests on Python 3.10/3.12.

These are transparent scenario diagnostics and heuristic scores, not financial forecasts or validated investment signals. Prompt Firewall is a diagnostic, not a security boundary. MIT applies to original code; the release introduces packaging and reproducibility rather than a new financial model.

Source commit: 257b85fcf0f585031bd0208f55e244f9220ced17.