Long-term

Release-independent research and optional distribution/integration proposals with no committed release date. Includes unresolved native generation, effective storage, handoff/ACK and first-input/startup acceptance (#180/#181/#268/#304/#306/#307), Codex side-session/steering (#245/#247), deduplication and optional integrations/distribution. Already shipped slices remain historical evidence, not proof that broader gates are complete. Preserve existing issue priorities, unsupported/unknown/no-automatic-resend boundaries and separate live-test/implementation/publication approvals.

Navigation Menu

Long-term

feat(distribution): offer the core CLI through an opt-in Homebrew tap

investigate(delivery): extend request-ID deduplication safely to ordinary local and SSH sends

investigate(integration): optional HOL Guard checkpoint for session-peer send and wake

investigate(integration): expose existing Codex sessions through Claude native peer tools

feat(experiment): validate opt-in Codex instant steering

feat(codex): investigate opt-in live steering with instant_interrupt

feat(experiment): validate opt-in Codex side-session transport

feat(codex): investigate opt-in delivery to live /btw side conversations

fix(identity): pin list-to-send targets to a session generation and fail closed when stale

feat(handoff): expose explicit ACK, deadline, and unknown outcome for unanswered requests

investigate(codex): align selected home inspection with native SQLite storage configuration

test(antigravity): validate CLI 1.3.3 on macOS and Linux before stable support

investigate(codex): receive input before the first turn through the owning App Server

investigate(antigravity): bootstrap agentapi delivery with a SessionStart hook

Issues