[Worflow] Make newton and rsl_rl default across core tasks

[StafaH]

Description

Important

Confirm the pull request base before submitting. Target develop for all
contributions. The release/3.0.0-beta2 branch is a frozen stable landing
snapshot and is not used for ongoing maintenance.

Please include a summary of the change and which issue is fixed. Please also include relevant motivation and context.
List any dependencies that are required for this change.

Fixes # (issue)

Type of change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (existing functionality will not work without user modification)
• Documentation update

Screenshots

Please attach before and after screenshots of the change if applicable.

Checklist

• I have read and understood the contribution guidelines
• I have run the pre-commit checks with ./isaaclab.sh --format
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• I have added a changelog fragment under source/<pkg>/changelog.d/ for every touched package (do not edit CHANGELOG.rst or bump extension.toml — CI handles that)
• I have added my name to the CONTRIBUTORS.md or my name already exists there

[greptile-apps]

Greptile Summary

The PR makes Newton MJWarp and RSL-RL the defaults across core tasks while preserving explicit backend selection.

• Adds task-registered default-agent dispatch when --rl_library is omitted.
• Changes core physics and renderer presets from Isaac Sim PhysX/RTX to kitless Newton defaults.
• Adds an RSL-RL PPO configuration for Pendulum and updates runtime/preset tests.

Confidence Score: 5/5

The PR appears safe to merge, with no concrete blocking or non-blocking defects identified in the changed behavior.

Default-agent lookup follows the repository's established Gym task-name convention, and the Newton physics and renderer defaults are coordinated with explicit alternatives and updated compatibility coverage.

Important Files Changed

Filename	Overview
source/isaaclab_rl/isaaclab_rl/entrypoints/dispatch.py	Resolves an omitted RL-library argument from the selected task's registered default agent while retaining selector help for tasks without a valid default.
source/isaaclab_tasks/isaaclab_tasks/utils/presets.py	Changes the default renderer to Newton Warp while retaining explicit Isaac RTX, automatic RTX, and OVRTX alternatives.
source/isaaclab_tasks/isaaclab_tasks/core/pendulum/agents/rsl_rl_ppo_cfg.py	Adds the RSL-RL PPO runner configuration needed by Pendulum's new default-agent registration.
source/isaaclab_tasks/isaaclab_tasks/core/lift/lift_env_cfg.py	Makes Newton MJWarp the default lift-task physics preset while preserving explicit PhysX alternatives.
source/isaaclab_tasks/test/core/test_runtime_compatibility.py	Updates compatibility expectations to verify the new Newton and kitless renderer defaults.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart LR
  CLI["Unified train/play CLI"] --> Task["Read --task"]
  Task --> Registry["Gym task registry"]
  Registry --> Agent["default_agent = rsl_rl"]
  Agent --> Backend["RSL-RL backend"]
  Registry --> Config["Core task configuration"]
  Config --> Physics["Newton MJWarp physics"]
  Config --> Renderer["Newton Warp renderer"]
  Physics --> Runtime["Kitless default runtime"]
  Renderer --> Runtime

Loading

_{Reviews (1): Last reviewed commit: "Make newton and rsl_rl default" | Re-trigger Greptile}

[isaaclab-review-bot]

Isaac Lab Review Bot

The new task-registered RSL-RL fallback and Newton MJWarp/Newton Warp defaults are implemented consistently, but both added changelog fragments are empty despite these user-visible behavior changes.

• Design and architecture: Task-level default_agent metadata provides a centralized backend-selection mechanism, while the preset aliases consistently establish Newton MJWarp and the Newton Warp renderer as defaults. No architectural issue was accepted beyond the missing release documentation for these decisions.
• API: Explicit --rl_library and explicit physics/renderer presets remain available, but omitted selections now resolve differently. These changed CLI and default-preset behaviors require package changelog entries with appropriate migration guidance.
• Implementation: The required fragment files exist for isaaclab_rl and isaaclab_tasks, but both are empty and therefore produce no compiled changelog entry. Populate them under the appropriate Changed category, marking the breaking default switch and documenting how users can explicitly retain prior behavior.

Minor fixes needed. Posted 1 actionable finding inline.

Automated review; human maintainers own approval decisions.

[isaaclab-review-bot]

🟡 Warning · Implementation — Empty changelog fragments for breaking default switch

This change flips resolved defaults (renderer to Newton Warp, core physics/decimation to newton_mjwarp) and makes --rl_library optional, but source/isaaclab_tasks/changelog.d/default-newton-rsl.minor.rst and source/isaaclab_rl/changelog.d/default-newton-rsl.minor.rst were added with no content, so the compiled changelog will carry no entry. Repository rules require a Changed entry per touched package, prefixed with **Breaking:**, plus migration guidance (e.g. selecting presets=isaacsim_physx,isaacsim_rtx to restore prior behavior).

[hujc7]

Looking great. Might still need to explicit resolve the physics for some tests if they were using the default physx?

[AntoineRichard]

Look, good, a lot of tasks change default from PhysX to Newton. If that's what we want then it's great!

[AntoineRichard]

I don't think the loc should have changed?

[kellyguo11]

should we keep all of our documentation and tutorials to use --viz newton by default with this change? and only use --viz kit when kit is required, and highlight in those docs that isaac sim is required for them

[ooctipus]

kwargs={
        "env_cfg_entry_point": f"{name}.joint_pos_env_cfg:FrankaCabinetEnvCfg",
        "agent_entry_point": preset(
                  default = f"{agents.name}.rsl_rl_ppo_cfg:CabinetPPORunnerCfg",
                  decrete_box = f"{agents.name}:skrl_manager_ppo_cfg.yaml"
                  rsl_rl_cfg_entry_point=f"{agents.name}.rsl_rl_ppo_cfg:CabinetPPORunnerCfg",
                  rl_games_cfg_entry_point = f"{agents.name}:rl_games_manager_ppo_cfg.yaml",
                  skrl_cfg_entry_point = f"{agents.name}:skrl_manager_ppo_cfg.yaml"
        )
    }
    
What do you think of option like this?

[StafaH]

Superseded by split PRs: #7064 (locomotion), #7065 (manipulation), #7066 (classic controls and renderer), #7067 (RSL-RL default), and this contrib coverage PR.

[StafaH]

The contrib coverage split is #7068.