What we intend to work on next, and where help is most valuable.
This document holds intent and priority only. It deliberately does not restate what currently works — that lives in one place, the Capability Matrix, with the outstanding per-platform items enumerated under Known Gaps.
Related: Cross-Platform Guide · Contributing · Architecture
Work that is not tied to any one platform:
- Strengthen macOS reliability around startup, config reloads, and mode transitions.
- Keep reducing global state to the minimum required by native bridge callbacks.
- Expand contract tests around ports, adapters, and reload behavior.
- Make unsupported platform capabilities fail loudly instead of silently no-oping.
Linux is beta and Windows alpha — meaning Linux is good for daily driving, while Windows is worth trying but not yet worth switching to. Every remaining item is tracked as a numbered entry in Known Gaps — pick one from there rather than from a duplicate list here, so the status you read is the status the code reports.
The two largest open areas:
- Linux — freedesktop notifications and alerts across all backends, and persisting the KDE RemoteDesktop portal grant across daemon restarts.
- Windows — foreground-window and display-hotplug events, which currently block per-app config re-application and monitor tracking.
GNOME Wayland remains unsupported; the daemon refuses to start there. Reviving it needs libei plus a GNOME Shell extension — see LINUX_DESKTOPS.md.
The highest-leverage areas, roughly in order:
- Platform adapter implementations in
internal/adapter/platform. - Overlay implementations and capability reporting.
- Config reload regression coverage.
- Reducing compatibility globals behind explicit interfaces.
New to the codebase? Contributing safely lists good starter tasks, the changes worth opening an issue for first, and the five-point bar a platform change has to clear before it lands.