Documentación de DIXI Spec Kit English · Código en GitHub

Esta página está disponible solo en inglés; el resumen en español, si existe, está al final.

B3 and v1.0.0 evidence

Purpose

Record the technical verification of the Claude Code adapter (TASK-003) and of version 1.0.0, and their limits. Date: 2026-11-08. Environment: macOS, Python 3.15.2, Claude Code 2.1.319, Setuptools 84.2.1 pinned by hash.

Tests first

BlockRed commitPrior result
Claude Code adapter[11-quality] test: acceptance in red for Claude Code adapter B318 expected failures; the --agent none regression passes
Public exporter[11-quality] test: acceptance in red for public release export3 expected failures
1.x compatibility[11-quality] test: accept same-major kit versions in red8 expected failures

Two frozen tests were changed with explicit authorization from the responsible person, each in its own commit: the quoting of the interpreter in the example's shim (paths with spaces) and the expected version (1.0.0.dev0 → 1.0.0).

Results

ArtifactBytesSHA-256
dixi_spec_kit-1.0.0-py3-none-any.whl53 42251e5a3424c6e1067b6bf2ba27e28d26228438390e46a44ae487c7a6c4f233a5e
dixi_spec_kit-1.0.0.tar.gz33 545ad736aa7096f977cf77af465e85afc757b790ca331dd5813455088be8ad35acf

The hashes identify this local build; bit-for-bit reproduction is not promised. The release must attach the artifacts built by the person from the published commit.

Walkthrough in the agent

The example was produced by following the steps of each command in a Claude Code 2.1.319 session, including the evolution with impact analysis and red tests before code (walkthrough). It is not an independent evaluation: a person still needs to invoke the ten installed /dixi.* commands in a new session and confirm that they behave according to the contract.

Publication

Pending