Public product state of falsegreen at a glance. For the full code catalog and usage,
see the README; for the change history, see the CHANGELOG.
Research artifacts, datasets, and unpublished numbers live in the private research hub, never in this repo. This file tracks the public product only.
- Current: 0.6.0 (PyPI:
pip install falsegreen) - Versioning: semver; releases via trusted publishing (OIDC).
ci.yml: tests on Python 3.8 / 3.11 / 3.13.release.yml: PyPI publish on tag.codex-review-gate.yml,release-drafter.yml,credit-contributor.yml.
Deterministic AST scan over pytest and unittest test code. Active codes:
- False-green (structural): C1, C2, C2b, C3, C4, C4b, C5, C6, C6b, C7, C8, C9, C11a, C13, C13b, C14, C16, C17, C18, C19, C20, C21, C22, C23, C24, C25, C27, C28, C29, C30, C31, C32, C33, C34, C35, C36, C37, C38, C39, C41, C42, C43, C44, C45, C48, CC.
- Diagnostic (opt-in, maintainability): D1, D3, D4, D5, D6.
- Coupling (opt-in): M2.
- Project layer (
--config-audit): PL2, PL7, PL8.
Each code carries a judgment tag (J1-J6) and a risk family (F1-F8); see the README catalog and the docs site for what each one flags, with a BAD plus CLEAN example.
Findings are tagged unit / integration / e2e by import roots. A real database or HTTP call in a unit test is itself the smell the level-awareness surfaces.
pytest and unittest (xUnit assert* forms), plus the common assertion and mock libraries.
Static layer only. Statically provable false-green with a low false-positive rate. Semantic
judgment goes to falsegreen-skill; runtime and culture are out of scope by design.