Merge remote-tracking branch 'origin/main' into robosuite-1to1

# Conflicts:
#	.gitignore

Author: geng-haoran

.gitignore

@@ -137,4 +137,9 @@ CLAUDE.local.md
 ENVIRONMENTS.local.md
 datasets/
 .robosuite_bundle/
+
+# --- LIBERO / LIBERO-plus 1:1 isolated integration ---
 third_party/
+INTEGRATION_LOG.md
+*.hdf5
+libero_datasets/

CHANGELOG.md

@@ -0,0 +1,102 @@
+# Changelog
+
+All notable changes to RoboVerse are documented in this file.
+
+The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
+and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
+
+RoboVerse is the downstream content-pack + research layer on top of the
+[MetaSim](https://github.com/RoboVerseOrg/MetaSim) core. See
+`MetaSim/CHANGELOG.md` for the core-framework changes that ship with this
+release.
+
+## [Unreleased]
+
+Nothing yet — open a PR to add entries here.
+
+## [1.0.0-beta] - 2026-05-31
+
+The headline of this release is **cross-platform parity is now testable and
+load-bearing**. Every shipped `RobotCfg`, every contracted handler method,
+and every benchmark `reset(seed=N)` is exercised across the supported
+backends and either passes or is xfail-documented with a specific reason.
+
+The release is forward- and backward-compatible.
+
+### Added
+
+- **`tests/test_roboverse_robot_cfg_validation.py`** — validates every
+  `RobotCfg` shipped in `roboverse_pack`: instantiation, non-empty `name`,
+  default joint positions inside `joint_limits` ranges. 60 cfgs × 4 checks,
+  9 real downstream bugs surfaced and xfail-documented:
+  - `AlohaAgilex`, `G1Tracking` — orphan joints in `default_joint_positions`
+  - `YamCfg`, `ArxL5Cfg`, `Vega`, `SoArm100`, `Koch`, `Go2`, `AllegroHand` —
+    defaults outside `joint_limits` ranges
+- **IL + RL fusion bridge** (`roboverse_learn/fusion/`) — RL-trained policy
+  → demo collection (rollout → `save_demo` → `data2zarr` unchanged) +
+  IL-to-RL BC warmstart. End-to-end pipeline orchestrator
+  (`rl-train → collect → to-zarr → il-train`). 6/6 tests; validated against
+  a real cartpole checkpoint and real mujoco/EGL launch. Fails fast on
+  scene-MJCF tasks where robot is not a `RobotCfg`.
+- **mjlab 1:1 obs/reward parity** — extensive byte-level alignment work:
+  - `TerrainGridScanSensor` + `height_scan` obs wired into
+    `velocity_rough_go1` / `velocity_rough_g1` (1:1 building block).
+  - Continuous rough terrain wired into both velocity-rough quadruped
+    tasks.
+  - `tracking-g1` motion-tracking obs ports `motion_anchor` + `robot_body`
+    1:1.
+  - `lift_cube_yam` regains the missing `ee_to_cube` + `cube_to_goal` obs.
+  - `velocity_rough` height_scan is now correctly scaled by
+    `1 / max_distance` (matches mjlab); clip-then-scale order on the obs
+    term matches mjlab's `noise → clip → scale`.
+  - go1 / g1 `base_lin_vel` obs reads the IMU velocimeter at the offset
+    site (adds the `ω × r` cross-term — fixes a real obs gap that diverged
+    under turning).
+  - cross-sim cartpole eval pipeline repaired (actor build, device,
+    episode metric).
+- **RoboTwin v2 passthrough** — 50 RoboTwin tasks registered as native gym
+  envs (`RoboVerse/RoboTwin-<task>-v0`), 1:1 by construction, mirrors the
+  ManiSkill passthrough pattern.
+- **`AGENTS.md` / `CLAUDE.md`** — repo-level dev rules for AI agents
+  (double quote, line-length 120, py38 target, lazy imports OK,
+  parity-is-load-bearing, multi-repo discipline, commit-as-user).
+
+### Fixed
+
+- **`get_started/0_static_scene.py`** — gracefully skips the cameras-empty
+  case instead of crashing on first-time setup.
+- **mjlab tooling paths** — lint-clean across mjlab/maniskill tasks;
+  tooling paths made portable (no hardcoded `/home/ghr/...`).
+- **mjlab MJCF asset resolution** — assets pulled from
+  `RoboVerseOrg/roboverse_data` on Hugging Face instead of hardcoded local
+  paths (fresh checkout no longer breaks). Dual-mode locator
+  `_locator.mjlab_asset` = clone-or-HF.
+- **mjlab go1 Newton 1:1** — go1 RL now stands on Newton (PD actuators +
+  init pose + action order alignment).
+
+### Documentation
+
+- `ROADMAP.md` — P0/P1 cross-platform infra items marked fixed; P2/P3
+  items refreshed.
+- mjlab / maniskill / robotwin integration pages wired into the docs
+  toctree.
+
+### Tests
+
+- 60 RoboVerse `RobotCfg` configs validated × 4 checks = 240 statically
+  exercised; 9 real bugs xfail-documented.
+- Cross-4-backend sweep against MetaSim core in the `roboverse` conda env:
+  mujoco + sapien3 + newton + passthroughs — 323 passed, 0 regressions.
+
+### Migration
+
+No code changes are required for existing users.
+
+If you maintain an out-of-tree `RobotCfg`, the new validation test will
+exercise it as soon as it is imported through `roboverse_pack` — fix any
+flagged `default_joint_positions` mismatches before they trip in CI.
+
+---
+
+[Unreleased]: https://github.com/RoboVerseOrg/RoboVerse/compare/v1.0.0-beta...HEAD
+[1.0.0-beta]: https://github.com/RoboVerseOrg/RoboVerse/compare/v1.0.0-alpha...v1.0.0-beta

RELEASE_NOTES_NEXT.md

@@ -0,0 +1,68 @@
+# RoboVerse — next release (paste-ready notes)
+
+**Suggested tag**: `v1.0.0-beta` (follows the `v1.0.0-alpha` lineage; the
+"beta" gate is the cross-platform-parity work landing in this release).
+
+**Usage**: rename `[Unreleased]` → `[1.0.0-beta] - 2026-05-31` in
+`CHANGELOG.md`, then:
+
+```bash
+git tag -a v1.0.0-beta -m "RoboVerse v1.0.0-beta — cross-platform parity is testable"
+git push origin v1.0.0-beta
+gh release create v1.0.0-beta --prerelease \
+  --title "RoboVerse v1.0.0-beta — cross-platform parity is testable" \
+  --notes-file RELEASE_NOTES_NEXT.md
+```
+
+This release should be cut alongside **MetaSim v0.2.0** (see
+`MetaSim/RELEASE_NOTES_NEXT.md`).
+
+---
+
+## Cross-platform parity is now load-bearing
+
+Every shipped `RobotCfg`, every contracted handler method, and every
+benchmark `reset(seed=N)` is exercised across the supported backends and
+either passes or is xfail-documented with a specific reason. The release is
+forward- and backward-compatible.
+
+### Highlights
+
+- **60 `RobotCfg`s validated**. `tests/test_roboverse_robot_cfg_validation.py`
+  exercises every shipped robot for instantiation, non-empty name, and
+  defaults-inside-limits. 9 real downstream bugs surfaced:
+  - `AlohaAgilex`, `G1Tracking` — orphan joints in `default_joint_positions`
+  - `YamCfg`, `ArxL5Cfg`, `Vega`, `SoArm100`, `Koch`, `Go2`, `AllegroHand` —
+    defaults outside `joint_limits` ranges
+- **mjlab 1:1 obs/reward parity** got measurably tighter. `velocity_rough_go1`
+  / `velocity_rough_g1` now ship `height_scan` obs (scale `1/max_distance`,
+  clip-then-scale order matches mjlab), continuous rough terrain, and
+  `tracking-g1` motion-tracking obs. go1/g1 `base_lin_vel` reads the IMU
+  velocimeter at the offset site (adds `ω × r` cross-term — closes a real
+  obs gap that diverged under turning).
+- **IL + RL fusion bridge** (`roboverse_learn/fusion/`) wires RL training →
+  demo collection → IL warmstart end-to-end. 6/6 tests, validated against a
+  real cartpole checkpoint and real mujoco/EGL launch.
+- **RoboTwin v2 passthrough** — 50 native gym envs registered, 1:1 by
+  construction (same pattern as the existing ManiSkill passthrough).
+- **`get_started/0_static_scene.py`** no longer crashes on first-time setup
+  when no cameras are configured.
+- **AGENTS.md / CLAUDE.md** lands: repo-level dev rules for AI agents
+  (parity is load-bearing, multi-repo discipline, commit-as-user).
+
+### Migration
+
+No required changes. If you maintain an out-of-tree `RobotCfg`, the new
+validation test will exercise it as soon as it is imported through
+`roboverse_pack` — fix any flagged `default_joint_positions` mismatches
+before they trip in CI.
+
+### Verification
+
+**323 tests pass** in the cross-4-backend sweep
+(mujoco + sapien3 + newton + passthroughs) in the `roboverse` conda env,
+zero regressions.
+
+### Full changelog
+
+See [`CHANGELOG.md`](./CHANGELOG.md) for the complete commit-by-commit list.

ROADMAP.md

@@ -15,16 +15,65 @@ This document provides a high-level overview of the RoboVerse improvement roadma
 
 ### P0 - Critical (Immediate)
 
-- [ ] **State Cache Consistency** - Fix data corruption when switching state formats
-- [ ] **Test Coverage** - Add validation tests for robot configs and core tasks
+- [x] **State Cache Consistency** - Fixed: dual tensor/dict caches in `BaseSimHandler`,
+  invalidated on `set_states` / `set_dof_targets` / `simulate`. Regression tests:
+  `metasim/test/sim/test_state_modes.py` (integration) + `metasim/test/test_set_states_key_validation.py`
+  (unit, no sim env).
+- [x] **`set_states` silent-drop guard** - Added: unknown / control-input keys now log a
+  one-shot warning (e.g. `dof_pos_target` → "use `set_dof_targets`"). Catches the silent
+  no-op bug that broke 15 downstream BC experiments.
+- [x] **Robot-config validation tests** - Added static `-k general` tests for every shipped
+  `RobotCfg`: instantiation, non-empty `name`, default joint positions match `joint_limits`
+  keys + lie inside the limit ranges. Surfaces 9 real config bugs (xfail-documented:
+  AlohaAgilex/G1Tracking orphan joints, YamCfg/ArxL5Cfg/Vega/SoArm100/Koch/Go2/AllegroHand
+  out-of-range defaults). `metasim/test/test_robot_cfg_validation_general.py` +
+  `tests/test_roboverse_robot_cfg_validation.py`.
+- [x] **Backend contract enforcement** - Static `-k general` test asserts every concrete
+  `BaseSimHandler` subclass overrides each documented contract method. Known incomplete
+  backends (pyrep, partial pybullet/genesis) are xfail-documented. Self-check guards the
+  xfail list against staleness. `metasim/test/test_backend_contract_general.py`.
 - [ ] **CI Coverage Reporting** - Integrate pytest-cov with Codecov
 
 ### P1 - High (Next Release)
 
-- [ ] **Abstract Method Declarations** - Complete @abstractmethod annotations
+- [x] **Parallel-sim error handling** - Fixed: every public method on `ParallelHandler`
+  now drains `error_queue` + checks `process.is_alive` after wire I/O. EOFError/BrokenPipe
+  on recv is translated into a `RuntimeError` carrying the real worker traceback instead
+  of a cryptic IPC exception. OOM-killed workers (queue stays empty) are surfaced via the
+  dead-process check. Tests: `metasim/test/test_parallel_error_handling_general.py`.
+- [x] **Newton joint-name parity** - Fixed: Newton's `_collect_joint_names` now strips
+  the object prefix (`box_base/box_joint` → `box_joint`) when the bare name is unambiguous,
+  matching mujoco / sapien3 / isaacgym / isaacsim. Cross-platform tasks doing
+  `dof_pos["box_joint"]` previously silently broke on Newton. Falls back to the full key
+  on collision so no information is lost. Verified by `test_dict_state_all_objects[newton-*]`
+  flipping from failed to passed.
+- [x] **Newton silent-action drop guard** - Added: Newton's `_set_dof_targets` now warns
+  once if `model.control()` has no `joint_target_pos` / `joint_target_vel` / `joint_f`
+  buffer — previously every action was silently swallowed. Same antipattern that broke
+  MuJoCo `set_states`.
+- [x] **Actuator partial-spec warning (MuJoCo + Newton)** - Identified as task #6 root
+  cause: when an actuator cfg overrides `stiffness` / `damping` but doesn't set
+  `effort_limit_sim`, the asset-authored force-range (MJCF `forcerange` / Newton
+  `joint_effort_limit`) silently dominates. Same `stiffness=1e5` config produces
+  different effective behaviour per backend. Both handlers now warn once per
+  (robot, joint) so the asymmetry is visible.
+- [x] **Benchmark reproducibility — `env.reset(seed=N)` now actually reseeds** -
+  End-to-end wired: `GymEnvAdapter.reset(seed=)` → `RLTaskEnv.reset(seed=)` /
+  `TaskBase.reset(seed=)` → `BaseSimHandler.set_seed`. Added a default
+  `set_seed` on the base handler that reseeds Python `random` + NumPy +
+  Torch (CPU and CUDA), so every backend gets reproducibility for free.
+  Backends with extra internal RNG (Newton warp kernels, Sapien physics
+  noise) can override and call `super().set_seed(seed)` first. Tests:
+  `test_set_seed_is_deterministic_on_numpy_and_torch` and
+  `test_set_seed_differs_with_different_seeds` in
+  `metasim/test/test_set_states_key_validation.py`. Backward-compat:
+  `seed` is a keyword-only parameter — every existing call site is
+  unchanged.
+- [ ] **Abstract Method Declarations** - Mostly covered by the new backend-contract test;
+  still want `@abstractmethod` on `_get_joint_names` / `_get_body_names` once pyrep /
+  partial pybullet / genesis catch up.
 - [ ] **Unified Environment Factory** - Create `roboverse.make_env()` API
 - [ ] **Configuration System** - Document and standardize config approaches
-- [ ] **Error Handling** - Improve parallel simulation error propagation
 
 ### P2 - Medium (Future)