v0.1.0

Initial release — a model-agnostic error-log and log-before-fix system for
AI coding agents.

Added

• Three-file system — rules.txt, errors.txt, notes.txt templates
  with placeholders for any project, plus example data that works out of the
  box (python start.py boots a fully working demo).
• check_errors.py tooling (stdlib only) — linter (template fields +
  canonical statuses FIXED | PARTIAL | OPEN | MITIGATED | WORKAROUND),
  --add scaffolder, --has-entry mechanical gate, --archive-days
  archiver (idempotent, dry-run by default), --log for any file.
• start.py session bootstrap — STEP 0 error-log health check, reading
  order, open errors, latest session note; configurable filenames.
• git-commitmsg-hook.sh — the log-before-fix git gate: blocks code
  commits whose message lacks an AREA: <text> marker matching a logged
  entry, lints the log when it is staged, and exposes env-var overrides
  (LOGNAME, PYTHON, AGENT_ERROR_LOG_DIR).
• _test_errors.py — 64 unit tests (parsing, validation, gate, add,
  archive, idempotency).
• CI — GitHub Actions workflow running the tests, linter, and hook
  syntax check on Linux and Windows across Python 3.9–3.12.
• Docs — README, AGENTS.md, CONTRIBUTING.md, MIT license.

Known limitations

• git commit --no-verify bypasses the local hook; documented with
  practical workarounds in the README.