Fixing agent mistakes

Fix each mistake at the strongest layer.

  1. Architecture

    Make the mistake impossible to write, for example by moving data access into one layer.

  2. Lint and CI

    A rule the linter or CI enforces on every PR, with no one watching.

  3. Review bot

    The AI reviewer flags it before a person reads the PR.

  4. Skill

    A skill teaches agents the right pattern for that kind of work.

  5. Doc

    The weakest layer. A rule left only as a PR comment is a bug.