MateProbe 0.1.0a4 — by Mate4B

MateProbe by Mate4B is the new name of Narrative Contracts. The project remains open source under MIT.

Install the core and optional pytest integration from the attached wheels:

python -m pip install \
  https://github.com/Mate4b/mateprobe/releases/download/v0.1.0a4/mateprobe-0.1.0a4-py3-none-any.whl \
  https://github.com/Mate4b/mateprobe/releases/download/v0.1.0a4/pytest_mateprobe-0.1.0a4-py3-none-any.whl

Python imports use mateprobe; the CLI is mateprobe; the pytest fixture is mateprobe, with --mateprobe-report for JSON output. Migration guide.

Contract behavior, report schema versions and frozen research evidence are unchanged. Stable historical rule identifiers preserve configuration digests. Old releases remain available under their original names.

Validation: 155 tests, Ruff/strict mypy, Python 3.11–3.14 CI, wheel/source installation and pytest/CLI/docs checks, 8/8 selected source mutants, and frozen replay equivalence allowing only runtime version metadata. These checks do not establish independent adoption or semantic accuracy.

The attached distributions were built from e842d66faa5d2dd6d8cc9a7dd8abd7e31bc49078; release-a4.json pins all four hashes. Both packages are published on PyPI; all four downloaded files match these release hashes. A clean PyPI installation passed the API, CLI, pytest, consumer documentation and launch regression checks.

python -m pip install mateprobe==0.1.0a4 pytest-mateprobe==0.1.0a4

Narrative Contracts 0.1.0a3 — audit existing validators

Audit an existing Python validator without adopting the Document/Claim model. Version 0.1.0a3 gives the independent audit APIs their own installable package version, synchronized across core, pytest plugin, and report metadata.

What is included

• Paired baseline/fault/control cases, finding-ID attribution, obligation inventories, explicit unknown/error outcomes, and CI thresholds.
• Reports prioritize survivors and other known gaps, incomplete evidence, and untested obligations. Independent evidence flags retain simultaneous gaps and execution failures; the display assessment remains available.
• Opt-in observed or supplied Git metadata, corpus digests, and versioned JSON reports.
• State-field bindings and bounded comparison/transition contracts.
• Progressive adoption from boolean validators, a survivor-to-regression pytest recipe, and executable JSON Schema/Pydantic configuration trials.

Install

python -m pip install narrative-contracts==0.1.0a3 pytest-narrative-contracts==0.1.0a3

The plugin is optional. Core runtime has no third-party dependencies and does not call a model or network during evaluation.

Verification and boundaries

Both wheels and source distributions were installed and exercised in separate clean environments. Validation includes 154 local tests with optional integrations, Ruff, mypy, documentation builds, selected source mutants, and frozen benchmark replay. Historical labels, findings and scores remain unchanged; replays permit only the explicit library-version change.

Integration cases and policies are authored. They demonstrate configuration gaps, not bugs in third-party libraries, independent adoption, production accuracy, or arbitrary-prose truthfulness. The refund prose-only survivor remains in the 8/9 denominator. Corpus design and label review remain the integrator's work.

Release artifacts were built from source commit 980c78f. The attached manifest and SHA256SUMS identify the exact bytes used for PyPI publication; subsequent documentation/publication metadata commits do not rebuild these files.

Documentation · Regression recipe · Source CI

Narrative Contracts 0.1.0a2 — reproducible open alpha

Narrative Contracts a2 adds reproducible evidence and a second application domain while keeping the core offline and dependency-free.

• Independent customer-support workflow: trusted transitions, branch selection, declared claims and valid paraphrases.
• Expanded synthetic campaign: 21 cases, 8/13 attributed fault detections, 4/5 controls preserved, and 3 explicit exclusions. Survivors and false positives remain public.
• Real-model pilot: all 32 responses from qwen3:1.7b and gemma3:1b, with frozen prompts/settings, raw outputs and deterministic offline replay.
• Community templates for missed obligations, valid text rejected and new contracts.
• Updated technical report. No human labels or LLM-as-judge labels were used; natural-output acceptance is not semantic accuracy.

Validation: 91 tests, Ruff, strict mypy, package builds, isolated wheel/plugin installation, eight selected source mutants killed, and offline replay. The attached provenance identifies the verified commit and CI run.

The pilot is deliberately published without score tuning: Qwen3 supplied 16 valid envelopes with consistent declarations but none passed the strict profile; all 16 Gemma3 responses failed the envelope. These results describe this small workload, not a general model ranking.

Install from the attached wheels:

python -m pip install \
  https://github.com/pablomate4b/narrative-contracts/releases/download/v0.1.0a2/narrative_contracts-0.1.0a2-py3-none-any.whl \
  https://github.com/pablomate4b/narrative-contracts/releases/download/v0.1.0a2/pytest_narrative_contracts-0.1.0a2-py3-none-any.whl

Core and plugin use MIT. Model weights are not distributed. The natural pilot zip includes model metadata, prompts and generated outputs; it is not human-approved advice. PyPI publication, independent annotation, external adoption and research submission remain optional follow-up work.

Technical report · Replay instructions · Submit a counterexample

v0.1.0a1 - Narrative Contracts Alpha

Narrative Contracts 0.1.0a1 (Alpha)

Primera alpha funcional de Narrative Contracts y su plugin de integración pytest-narrative-contracts.

Contenido de la versión:

• Núcleo Python sin dependencias externas con 8 tipos de contratos deterministas y auditoría de mutaciones para generación de lenguaje condicionada por estado.
• Plugin pytest-narrative-contracts, CLI y reportes reproducibles estructurados con evidencia.
• Atribución de hallazgos por regla/código/ubicación, controles positivos y exclusiones visibles.
• Adaptador y ejemplo de integración con LifeCard.
• Corpus sintético de factibilidad (720 casos: 80% detección de defectos y 80% preservación de variantes válidas) y suite de regresión con detección de 8/8 mutaciones de código seleccionadas.
• Borrador de paper de factibilidad con protocolo para evaluación empírica independiente sobre salidas reales.

Artefactos adjuntos:

• narrative_contracts-0.1.0a1-py3-none-any.whl
• narrative_contracts-0.1.0a1.tar.gz
• pytest_narrative_contracts-0.1.0a1-py3-none-any.whl
• pytest_narrative_contracts-0.1.0a1.tar.gz

Nota: La publicación en PyPI queda programada para una etapa posterior tras la evaluación independiente con salidas reales.