v0.2.0 — Practices Auditor

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog,
and this project adheres to Semantic Versioning.

[0.2.0] - 2026-04-27

Added

• decipher practices command -- repository health auditor that scores a
  Python project against software-development best practices and produces a
  structured report with per-category scores and prioritised recommendations.
• 8 practice checkers covering project structure (PROJ-001..011), testing
  (TEST-001..005), quality gates (QUAL-001..005), CI/CD (CICD-001..006),
  dependency hygiene (DEP-001..005), documentation (DOC-001..005), licensing
  (LIC-001..003), and release readiness (REL-001..007).
• Finding registry with compact IDs (PREFIX-NNN) and gate codes (DEP-001,
  DOC-001) that conditionally suppress downstream findings when a prerequisite
  check fails.
• Severity model -- every finding is fail (can't ship) or warn (should
  improve). Scoring formula: max(0, 100 - warn*10 - fail*25).
• CLI flags for the practices command:
  • --format terminal|markdown|json (auto-detected from TTY, CI environment
    variables, or -o file extension).
  • --strict -- exit 1 on warnings, not just failures.
  • --only <checkers> / --skip <checkers> -- run or exclude specific
    checkers (comma-separated, mutually exclusive).
  • -o <path> -- write report to file; format inferred from extension
    (.json, .md).
• Exit-code semantics -- 0 (pass, or warn without --strict), 1 (fail, or
  warn with --strict), 2 (input/argument error).
• Dogfood gate -- the tool audits itself (decipher practices .) and must
  score 100/100 before shipping.
• CI workflow (.github/workflows/ci.yml) -- pytest across Python
  3.10--3.13, ruff lint, and mypy (informational, continue-on-error: true).
• Coverage configuration ([tool.coverage.run]/[tool.coverage.report])
  with fail_under = 80 and an omit list for LLM-dependent modules.
• mypy strict configuration ([tool.mypy] strict = true) in pyproject.toml.
• Pre-commit hooks -- ruff lint + format, plus basic hygiene hooks
  (trailing-whitespace, end-of-file-fixer, check-yaml, check-toml,
  check-merge-conflict).
• Reporter with three output backends: to_json(), to_markdown(), and
  to_terminal() (Rich Panel + Table + Text, returned as
  rich.console.Group).

Changed

• License changed from Apache-2.0 to MIT. See LICENSE and
  MIGRATION.md for details.
• Reporter.to_terminal() now returns rich.console.Group instead of printing
  directly. Downstream Python consumers that called this method must update
  their rendering code. CLI users are unaffected.

Deprecated

• --json-only flag on decipher practices. Use --format json instead.
  The old flag is retained as a hidden alias for one release cycle and will be
  removed in v0.3.

Fixed

• Added TYPE_CHECKING guard imports in cli.py, resolving 3 mypy
  [name-defined] errors for LLMClient and AnalysisResult.
• Renamed ambiguous variable l to ln in analyzer.py and interactive.py
  (ruff E741).

Internal

• Phase-gated development pattern (A-F) introduced for v0.2.0; used as
  governance baseline for future releases.

Known Limitations (v0.3 backlog)

• 45 mypy strict-mode errors remain (approximately 30 in practices checkers,
  15 in legacy modules). The CI mypy step is informational
  (continue-on-error: true) and will be promoted to blocking once errors
  reach zero.
• LLM-dependent modules (archaeologist, diagrammer, interactive,
  readme_generator) are excluded from coverage measurement. Adding response
  recorders (e.g. vcrpy) is planned for v0.3.
• --json-only hidden alias will be removed in v0.3.
• Only Python repositories are supported. Additional language support is
  planned for future releases.

[0.1.0] - 2026-04-17

Added

• Initial release of DecipherCode
• Full codebase analysis with LLM-powered architecture detection
• README generation from any codebase
• Architecture diagrams (Mermaid, GraphViz DOT)
• Git archaeology reports with contributor and hotspot analysis
• Interactive Q&A mode for natural language codebase exploration
• Multi-provider LLM support:
  • OpenAI
  • Azure OpenAI
  • Anthropic (via proxy)
  • Ollama (local)
• Support for local directories and GitHub URLs as input
• JSON export for scan results

v0.1.0 - Initial Release

First public release of DecipherCode. See CHANGELOG.md for details.