A2A routing, chat landing, env config, and site SEO.
-
Added A2A routing and generic metadata support on messages, expanded JSON-RPC compatibility aliases, and covered the behavior in API docs, tests, and smoke checks.
-
Added authenticated human task dispatch through the v1 API with OpenAPI and usage documentation.
-
Hardened expired invitation identity URIs and S3 state handling while expanding auth, command JSON, handle, router, and store test coverage.
-
Shared library agent instructions into packaged runtime contexts so seeded agents receive the repository guidance they need during task execution.
-
Applied environment variables to harness and runtime application configuration, with config-script coverage for persisted settings.
-
Simplified task-list GitHub status display so only one GitHub action appears at the final stage.
-
Kept the history view stable when submitting tasks and centralized auth/runtime configuration helpers for harness and Hub API setup.
-
Added a chat-first agent landing page with route, Dock, auth callback, and session-storage updates so app entry can open directly into agent chat.
-
Preserved login query parameters through auth callback redirects and aligned the chat page title format with the rest of the app.
-
Introduced a bolder shared Molten brand header across dashboard, settings, and Hub management surfaces.
-
Refined settings and Hub connection prompt copy with regression coverage for the updated onboarding language.
-
Removed the blocking external font dependency and added font-loading regression coverage for faster, more reliable page rendering.
-
Published the April 27 release rollup and improved release project links so cards and detail pages route to the right product destinations.
-
Expanded meta descriptions, redirects, Hub region content, and validation tests across blog, knowledge-base, release, and Hub pages.
-
Tightened homepage OpenClaw copy, refreshed task screenshots, and compacted Hub entity stats with clearer icon-driven presentation.
-
Added a default skill selection when choosing an agent, with server and UI coverage for the skill picker flow.
-
Applied deployment environment variables to Dispatch application configuration and documented the updated startup behavior.