Files

T

Luke The Dev 65c2b9cf85 Avatar Customization + Update Agent Brain (#23 )

Co-authored-by: iamlukethedev <iamlukethedev@users.noreply.github.com>

2026-03-20 23:05:14 -05:00

Changelog

[0.1.2] - 2026-03-20

An in-app avatar creator for agents with live 3D preview, appearance presets, and accessory controls for customizing office avatars.
A unified agent editor modal in the office that lets you edit avatars alongside agent brain files such as IDENTITY.md, SOUL.md, AGENTS.md, USER.md, TOOLS.md, MEMORY.md, and HEARTBEAT.md.
Structured avatar profile persistence and normalization so studio settings can store full avatar appearance data per gateway and agent instead of only avatar seeds.
A DEBUG environment toggle for controlling the OpenClaw event console in the office UI.

Reworked office avatar rendering so 3D agents reflect saved appearance profiles, including hair, clothing, hats, glasses, headsets, backpacks, and other visual variations.
Replaced avatar shuffle entry points in the chat and office surfaces with avatar customization flows that open the editor directly.
Updated the office HUD with a compact agent roster, overflow handling, and direct shortcuts into per-agent editing from the 3D office view.
Expanded the brain editor so IDENTITY.md fields are edited in structured form and agent renames can be applied to the live gateway agent after saving.
Defaulted the OpenClaw event console to a collapsed state and made it optional from environment configuration.
Updated hydration and store state to carry full avatar profiles through agent loading, persistence, and rendering.

Fixed WebSocket gateway authentication during the upgrade handshake by wiring access control through the ws verifyClient flow.
Fixed the gym release directive TypeScript error by adding explicit "release" support to office gym directives and aligning release-hold logic.
Corrected studio settings merging and normalization for avatar data so saved office appearances survive reloads and patch updates.
Kept skill gym hold state active for release directives during office animation trigger reconciliation.

Added unit coverage for avatar profile persistence, studio settings normalization, and fleet hydration with structured avatar data.
Expanded end-to-end coverage for avatar settings fixtures, office header and sidebar flows, voice reply settings persistence, disconnected office settings surfaces, and office route expectations.

Initial public Claw3D project documentation, including README.md, VISION.md, and ARCHITECTURE.md.
A gateway-first web UI for connecting to OpenClaw agents, monitoring runtime activity, and managing agent workflows.
A retro-office 3D environment for visualizing agent activity, spatial interactions, and immersive operational surfaces.
An office builder flow for editing and publishing office layouts.