Why:
- The approved grouping design needs an executable sequence that preserves settings behavior and existing user changes.
What:
- Define the TDD steps, exact catalog order, README copy, selective commits, and regression checks.
Why:
- The settings screen and README use inconsistent categories, which makes options harder to find.
What:
- Define the approved six-group taxonomy, compatibility constraints, and test strategy.
Why:
- The approved architecture design needs an executable task-by-task plan before implementation begins.
- A committed plan gives subagents a stable requirements source and preserves recovery context across long-running work.
What:
- Add a staged TDD implementation plan for candidate scanning, formatting-run extraction, settings catalog, Markdown segmentation, and URL formatting orchestration.
- Include exact files, interfaces, test commands, verification steps, and commit points for each stage.
Why:
- The architecture deepening work will span multiple core modules, so the design needs to remove vague implementation guidance before planning begins.
- Clearer scope reduces the chance that future agents move the wrong tests or leave settings and Markdown segmentation decisions ambiguous.
What:
- Specify that pure transformation tests should move from main adapter tests into formatting-run tests.
- Define the URL formatting settings projection in concrete terms.
- Make the Markdown segment extraction order explicit for the implementation plan.
Why:
The architecture review identified five related deepening opportunities that should be implemented in a staged, testable sequence rather than as ad hoc refactors.
What:
Document the approved design for candidate scanning, formatting runs, settings catalog, Markdown protected segments, and URL formatting adapters, including scope, non-goals, testing strategy, migration order, risks, and success criteria.
