eduardstan

eduardstan/reasonsmith

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

2 stars 1 forks 2 watchers Python MIT License

ai-governance ai-safety auditing compliance conformance-testing eu-ai-act explainable-ai formal-methods gdpr interpretability legaltech model-checking neurosymbolic python regtech responsible-ai smt-solver temporal-logic trustworthy-ai z3

View on GitHub Website

24 Open Issues Need Help Last updated: Aug 2, 2026

Open Issues Need Help

View All on GitHub

notebook: the report as the person the decision was about sees it about 1 month ago

documentation help wanted

eduardstan/reasonsmith

2

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

Python

#ai-governance#ai-safety#auditing#compliance#conformance-testing#eu-ai-act#explainable-ai#formal-methods#gdpr#interpretability#legaltech#model-checking#neurosymbolic#python#regtech#responsible-ai#smt-solver#temporal-logic#trustworthy-ai#z3

notebook: the auditor's walk-through — what to ask for, and what to refuse about 1 month ago

documentation good first issue

eduardstan/reasonsmith

2

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

Python

#ai-governance#ai-safety#auditing#compliance#conformance-testing#eu-ai-act#explainable-ai#formal-methods#gdpr#interpretability#legaltech#model-checking#neurosymbolic#python#regtech#responsible-ai#smt-solver#temporal-logic#trustworthy-ai#z3

docs: 51 cross-references point at sections `docs/semantics.md` does not have about 1 month ago

documentation good first issue

eduardstan/reasonsmith

2

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

Python

#ai-governance#ai-safety#auditing#compliance#conformance-testing#eu-ai-act#explainable-ai#formal-methods#gdpr#interpretability#legaltech#model-checking#neurosymbolic#python#regtech#responsible-ai#smt-solver#temporal-logic#trustworthy-ai#z3

feat(examples): a worked example wrapping a model from a mainstream ML library about 1 month ago

help wanted good first issue wanted: model/SUT

eduardstan/reasonsmith

2

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

Python

#ai-governance#ai-safety#auditing#compliance#conformance-testing#eu-ai-act#explainable-ai#formal-methods#gdpr#interpretability#legaltech#model-checking#neurosymbolic#python#regtech#responsible-ai#smt-solver#temporal-logic#trustworthy-ai#z3

feat(pack): bring one of the queued regulation candidates through the gates about 1 month ago

help wanted wanted: pack

eduardstan/reasonsmith

2

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

Python

#ai-governance#ai-safety#auditing#compliance#conformance-testing#eu-ai-act#explainable-ai#formal-methods#gdpr#interpretability#legaltech#model-checking#neurosymbolic#python#regtech#responsible-ai#smt-solver#temporal-logic#trustworthy-ai#z3

test: a clean install should give a green suite — skip loudly when an optional extra is absent about 1 month ago

help wanted good first issue

eduardstan/reasonsmith

2

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

Python

#ai-governance#ai-safety#auditing#compliance#conformance-testing#eu-ai-act#explainable-ai#formal-methods#gdpr#interpretability#legaltech#model-checking#neurosymbolic#python#regtech#responsible-ai#smt-solver#temporal-logic#trustworthy-ai#z3

docs(crosswalk): map a third assurance framework onto the evidence machinery about 1 month ago

documentation help wanted

eduardstan/reasonsmith

2

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

Python

#ai-governance#ai-safety#auditing#compliance#conformance-testing#eu-ai-act#explainable-ai#formal-methods#gdpr#interpretability#legaltech#model-checking#neurosymbolic#python#regtech#responsible-ai#smt-solver#temporal-logic#trustworthy-ai#z3

feat(pack): AI Act Article 50 transparency — the duty in force, covered by no pack about 1 month ago

help wanted wanted: pack

eduardstan/reasonsmith

2

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

Python

#ai-governance#ai-safety#auditing#compliance#conformance-testing#eu-ai-act#explainable-ai#formal-methods#gdpr#interpretability#legaltech#model-checking#neurosymbolic#python#regtech#responsible-ai#smt-solver#temporal-logic#trustworthy-ai#z3

feat(adapters): a second decision-log format beside JSONL about 1 month ago

enhancement help wanted good first issue

eduardstan/reasonsmith

2

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

Python

#ai-governance#ai-safety#auditing#compliance#conformance-testing#eu-ai-act#explainable-ai#formal-methods#gdpr#interpretability#legaltech#model-checking#neurosymbolic#python#regtech#responsible-ai#smt-solver#temporal-logic#trustworthy-ai#z3

docs(roadmap): the static registry shipped but the roadmap still lists it as open about 1 month ago

documentation good first issue

eduardstan/reasonsmith

2

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

Python

#ai-governance#ai-safety#auditing#compliance#conformance-testing#eu-ai-act#explainable-ai#formal-methods#gdpr#interpretability#legaltech#model-checking#neurosymbolic#python#regtech#responsible-ai#smt-solver#temporal-logic#trustworthy-ai#z3

feat(assessment): land a source-backed open-textured duty about 1 month ago

enhancement help wanted

eduardstan/reasonsmith

2

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

Python

#ai-governance#ai-safety#auditing#compliance#conformance-testing#eu-ai-act#explainable-ai#formal-methods#gdpr#interpretability#legaltech#model-checking#neurosymbolic#python#regtech#responsible-ai#smt-solver#temporal-logic#trustworthy-ai#z3

feat(plugins): validate external witnesses in the trusted core about 1 month ago

enhancement help wanted

eduardstan/reasonsmith

2

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

Python

#ai-governance#ai-safety#auditing#compliance#conformance-testing#eu-ai-act#explainable-ai#formal-methods#gdpr#interpretability#legaltech#model-checking#neurosymbolic#python#regtech#responsible-ai#smt-solver#temporal-logic#trustworthy-ai#z3

feat(docs): publish a static installed-plugin registry about 1 month ago

enhancement help wanted

eduardstan/reasonsmith

2

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

Python

#ai-governance#ai-safety#auditing#compliance#conformance-testing#eu-ai-act#explainable-ai#formal-methods#gdpr#interpretability#legaltech#model-checking#neurosymbolic#python#regtech#responsible-ai#smt-solver#temporal-logic#trustworthy-ai#z3

feat(observed): widen temporal monitor rule sets about 1 month ago

enhancement good first issue

eduardstan/reasonsmith

2

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

Python

#ai-governance#ai-safety#auditing#compliance#conformance-testing#eu-ai-act#explainable-ai#formal-methods#gdpr#interpretability#legaltech#model-checking#neurosymbolic#python#regtech#responsible-ai#smt-solver#temporal-logic#trustworthy-ai#z3

docs(cli): document the installed scaffold path end to end about 1 month ago

documentation good first issue

eduardstan/reasonsmith

2

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

Python

#ai-governance#ai-safety#auditing#compliance#conformance-testing#eu-ai-act#explainable-ai#formal-methods#gdpr#interpretability#legaltech#model-checking#neurosymbolic#python#regtech#responsible-ai#smt-solver#temporal-logic#trustworthy-ai#z3

docs(tui): no documented way to install or launch the terminal interface about 1 month ago

documentation help wanted good first issue

eduardstan/reasonsmith

2

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

Python

#ai-governance#ai-safety#auditing#compliance#conformance-testing#eu-ai-act#explainable-ai#formal-methods#gdpr#interpretability#legaltech#model-checking#neurosymbolic#python#regtech#responsible-ai#smt-solver#temporal-logic#trustworthy-ai#z3

fix(cli): launcher help advertises list-packs and serve, which do not exist about 1 month ago

bug help wanted good first issue

eduardstan/reasonsmith

2

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

Python

#ai-governance#ai-safety#auditing#compliance#conformance-testing#eu-ai-act#explainable-ai#formal-methods#gdpr#interpretability#legaltech#model-checking#neurosymbolic#python#regtech#responsible-ai#smt-solver#temporal-logic#trustworthy-ai#z3

feat(examples): add an ONNX artifact-backed neural SUT about 1 month ago

help wanted wanted: model/SUT

eduardstan/reasonsmith

2

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

Python

#ai-governance#ai-safety#auditing#compliance#conformance-testing#eu-ai-act#explainable-ai#formal-methods#gdpr#interpretability#legaltech#model-checking#neurosymbolic#python#regtech#responsible-ai#smt-solver#temporal-logic#trustworthy-ai#z3

feat(tui): make evidence rungs and refusals legible at a glance about 1 month ago

help wanted good first issue wanted: design

eduardstan/reasonsmith

2

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

Python

#ai-governance#ai-safety#auditing#compliance#conformance-testing#eu-ai-act#explainable-ai#formal-methods#gdpr#interpretability#legaltech#model-checking#neurosymbolic#python#regtech#responsible-ai#smt-solver#temporal-logic#trustworthy-ai#z3

feat(neural): admit alpha-beta-CROWN at proved about 1 month ago

help wanted wanted: engine

eduardstan/reasonsmith

2

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

Python

#ai-governance#ai-safety#auditing#compliance#conformance-testing#eu-ai-act#explainable-ai#formal-methods#gdpr#interpretability#legaltech#model-checking#neurosymbolic#python#regtech#responsible-ai#smt-solver#temporal-logic#trustworthy-ai#z3

feat(neural): raise a neural SUT above observed about 1 month ago

enhancement help wanted

eduardstan/reasonsmith

2

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

Python

#ai-governance#ai-safety#auditing#compliance#conformance-testing#eu-ai-act#explainable-ai#formal-methods#gdpr#interpretability#legaltech#model-checking#neurosymbolic#python#regtech#responsible-ai#smt-solver#temporal-logic#trustworthy-ai#z3

docs: the new-issue chooser does not route questions to Discussions about 2 months ago

documentation good first issue

eduardstan/reasonsmith

2

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

Python

#ai-governance#ai-safety#auditing#compliance#conformance-testing#eu-ai-act#explainable-ai#formal-methods#gdpr#interpretability#legaltech#model-checking#neurosymbolic#python#regtech#responsible-ai#smt-solver#temporal-logic#trustworthy-ai#z3

build: pack proposals arrive unlabelled — the template never applies the pack label about 2 months ago

good first issue pack

eduardstan/reasonsmith

2

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

Python

#ai-governance#ai-safety#auditing#compliance#conformance-testing#eu-ai-act#explainable-ai#formal-methods#gdpr#interpretability#legaltech#model-checking#neurosymbolic#python#regtech#responsible-ai#smt-solver#temporal-logic#trustworthy-ai#z3

docs: the documentation index points at a file that was renamed, and misses two documents about 2 months ago

documentation good first issue

eduardstan/reasonsmith

2

Check whether a system actually respects the legal requirements it is subject to — with verdicts that state how they were reached: proved, probed, observed, or unattainable.

Python

#ai-governance#ai-safety#auditing#compliance#conformance-testing#eu-ai-act#explainable-ai#formal-methods#gdpr#interpretability#legaltech#model-checking#neurosymbolic#python#regtech#responsible-ai#smt-solver#temporal-logic#trustworthy-ai#z3