Skip to content

CHANGELOG.md

Latest commit

 

History

History
67 lines (49 loc) · 4.29 KB

CHANGELOG.md

File metadata and controls

67 lines (49 loc) · 4.29 KB

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.

[1.2.1] - 2026-03-25

Fixed

  • Level-5-autonomous fixture: added missing deps (express, @types/express, supertest), fixed TypeScript implicit any, added posttest caching script
  • Readiness skill now checks both .git/hooks/ and .husky/ for hook-related criteria

[1.2.0] - 2026-03-25

Added

  • 86 unit tests for enforcement scripts: check-file-sizes, check-secrets, check-test-colocation, validate-docs, and incremental doc indexing
  • 23 E2E tests for post-setup lifecycle: enforcement round-trips against real git repos, doc indexing subprocess, git commit hook integration, doc drift detection
  • Python/FastAPI adaptive path eval scenario (setup-python)
  • Setup-then-readiness loop eval — verifies /setup output scores Level 3+ on /readiness
  • Multi-step eval support in run-evals.sh (steps array in test case config)
  • Hook-commit validation in setup grader: simulates real git commit after setup
  • Deny list validation in setup grader: verifies rm -rf, push --force, reset --hard are blocked
  • CLAUDE.md content checks in setup grader: claude_md_must_mention / claude_md_must_not_mention

Fixed

  • Glob matching bug: **/ now matches zero or more directories. Previously src/**/*.js silently skipped files directly in src/ — only matched files in subdirectories. Affected check-file-sizes, check-secrets, and check-test-colocation.
  • Pre-commit hook missing set -e: enforcement script failures (e.g. secret detection exit 1) were silently ignored and commits went through anyway
  • Readiness report path: moved from .claude/readiness-report.md to project root readiness-report.md — .claude/ is write-protected in non-interactive modes
  • Eval permissions: seed .claude/settings.json with tool permissions before git init, disable GPG signing for eval commits
  • Eval hook detection: grader now checks both .git/hooks/ and .husky/ for hook existence and executability
  • Eval timeouts: readiness 300→600s, setup 600→900s, setup-then-readiness 1200s

[1.1.0] - 2026-03-24

Added

  • /setup skill verification phase (Phase 6): 6 smoke checks run after installation to confirm hooks, enforcement scripts, CLAUDE.md, agent config, auto-docs pipeline, and linter all work
  • Setup eval suite: setup-eval-config.json, setup-grader.js, and 2 fixtures (setup-bare, setup-existing-node)
  • run-evals.sh now accepts --config flag and per-test-case prompt overrides
  • scripts/repo-generate-docs.js: auto-regenerates CLAUDE.md AUTO:tree and AUTO:modules markers on every commit via pre-commit hook
  • OSS infrastructure: CI workflow, issue/PR templates, CODEOWNERS, CONTRIBUTING, SECURITY, CODE_OF_CONDUCT, CHANGELOG, .editorconfig

Fixed

  • extractJSDocDescription now tries multi-line JSDoc before single-line (was picking up @param annotations instead of file-level descriptions)
  • Pre-push hook test cache: added missing exit 0 after cache hit (was printing "skipping" but running anyway)
  • Added results to SKIP_DIRS so eval output directories are excluded from generated trees
  • Added .env, coverage/, dist/, build/ to .gitignore

[1.0.0] - 2026-03-24

Added

  • /readiness skill: evaluates codebases across 8 pillars and 5 maturity levels with scored reports
  • /setup skill: scaffolds CLAUDE.md files, enforcement scripts, git hooks, linter configs, and agent settings
  • Enforcement scripts: secret scanning, file size limits, test colocation, doc generation, drift detection
  • Git hook templates: pre-commit (6 checks) and pre-push (test suite with SHA caching)
  • CLAUDE.md templates: global (cross-project standards) and project (per-codebase guidance)
  • Path-scoped rule templates: TDD, code quality, testing, TypeScript
  • Config templates: ESLint, Prettier, lint-staged, .gitignore, .env.example, Claude Code settings
  • Reference mapping of 20+ best practices from OpenAI, Anthropic, Augment Code, Factory.ai, and practitioners
  • Evaluation suite with 3 fixture projects (level 1, 3, 5) and automated grading
  • Unit tests for all setup scripts
  • Plugin manifest for Claude Code marketplace