Removes pickle dependency for cfg load and dump (isaac-sim#3709)

# Description

We have been supporting both pickle and yaml storing for configuration.
However, pickle has some security vulnerabilities and we have been
preferring the use of yaml in most cases. Thus, we are removing the
pickle utilities for saving and loading configs.

For more info on pickle: https://docs.python.org/3/library/pickle.html


## Type of change

- Breaking change (existing functionality will not work without user
modification)


## Checklist

- [x] I have read and understood the [contribution
guidelines](https://isaac-sim.github.io/IsaacLab/main/source/refs/contributing.html)
- [x] I have run the [`pre-commit` checks](https://pre-commit.com/) with
`./isaaclab.sh --format`
- [ ] I have made corresponding changes to the documentation
- [x] My changes generate no new warnings
- [ ] I have added tests that prove my fix is effective or that my
feature works
- [x] I have updated the changelog and the corresponding version in the
extension's `config/extension.toml` file
- [x] I have added my name to the `CONTRIBUTORS.md` or my name already
exists there

---------

Signed-off-by: Kelly Guo <kellyg@nvidia.com>
Co-authored-by: Mayank Mittal <12863862+Mayankm96@users.noreply.github.com>

Author: 2 people

scripts/benchmarks/benchmark_rlgames.py

@@ -75,7 +75,7 @@
 
 from isaaclab.envs import DirectMARLEnvCfg, DirectRLEnvCfg, ManagerBasedRLEnvCfg
 from isaaclab.utils.dict import print_dict
-from isaaclab.utils.io import dump_pickle, dump_yaml
+from isaaclab.utils.io import dump_yaml
 
 from isaaclab_rl.rl_games import RlGamesGpuEnv, RlGamesVecEnvWrapper
 
@@ -168,8 +168,6 @@ def main(env_cfg: ManagerBasedRLEnvCfg | DirectRLEnvCfg | DirectMARLEnvCfg, agen
     # dump the configuration into log-directory
     dump_yaml(os.path.join(log_root_path, log_dir, "params", "env.yaml"), env_cfg)
     dump_yaml(os.path.join(log_root_path, log_dir, "params", "agent.yaml"), agent_cfg)
-    dump_pickle(os.path.join(log_root_path, log_dir, "params", "env.pkl"), env_cfg)
-    dump_pickle(os.path.join(log_root_path, log_dir, "params", "agent.pkl"), agent_cfg)
 
     # read configurations about the agent-training
     rl_device = agent_cfg["params"]["config"]["device"]

scripts/benchmarks/benchmark_rsl_rl.py

@@ -75,7 +75,7 @@
 
 from isaaclab.envs import DirectMARLEnvCfg, DirectRLEnvCfg, ManagerBasedRLEnvCfg
 from isaaclab.utils.dict import print_dict
-from isaaclab.utils.io import dump_pickle, dump_yaml
+from isaaclab.utils.io import dump_yaml
 
 from isaaclab_rl.rsl_rl import RslRlOnPolicyRunnerCfg, RslRlVecEnvWrapper
 
@@ -207,8 +207,6 @@ def main(env_cfg: ManagerBasedRLEnvCfg | DirectRLEnvCfg | DirectMARLEnvCfg, agen
     # dump the configuration into log-directory
     dump_yaml(os.path.join(log_dir, "params", "env.yaml"), env_cfg)
     dump_yaml(os.path.join(log_dir, "params", "agent.yaml"), agent_cfg)
-    dump_pickle(os.path.join(log_dir, "params", "env.pkl"), env_cfg)
-    dump_pickle(os.path.join(log_dir, "params", "agent.pkl"), agent_cfg)
 
     benchmark.set_phase("sim_runtime")
 

scripts/reinforcement_learning/rl_games/train.py

@@ -79,7 +79,7 @@
 )
 from isaaclab.utils.assets import retrieve_file_path
 from isaaclab.utils.dict import print_dict
-from isaaclab.utils.io import dump_pickle, dump_yaml
+from isaaclab.utils.io import dump_yaml
 
 from isaaclab_rl.rl_games import MultiObserver, PbtAlgoObserver, RlGamesGpuEnv, RlGamesVecEnvWrapper
 
@@ -146,8 +146,6 @@ def main(env_cfg: ManagerBasedRLEnvCfg | DirectRLEnvCfg | DirectMARLEnvCfg, agen
     # dump the configuration into log-directory
     dump_yaml(os.path.join(log_root_path, log_dir, "params", "env.yaml"), env_cfg)
     dump_yaml(os.path.join(log_root_path, log_dir, "params", "agent.yaml"), agent_cfg)
-    dump_pickle(os.path.join(log_root_path, log_dir, "params", "env.pkl"), env_cfg)
-    dump_pickle(os.path.join(log_root_path, log_dir, "params", "agent.pkl"), agent_cfg)
 
     # read configurations about the agent-training
     rl_device = agent_cfg["params"]["config"]["device"]

scripts/reinforcement_learning/rsl_rl/train.py

@@ -88,7 +88,7 @@
     multi_agent_to_single_agent,
 )
 from isaaclab.utils.dict import print_dict
-from isaaclab.utils.io import dump_pickle, dump_yaml
+from isaaclab.utils.io import dump_yaml
 
 from isaaclab_rl.rsl_rl import RslRlBaseRunnerCfg, RslRlVecEnvWrapper
 
@@ -196,8 +196,6 @@ def main(env_cfg: ManagerBasedRLEnvCfg | DirectRLEnvCfg | DirectMARLEnvCfg, agen
     # dump the configuration into log-directory
     dump_yaml(os.path.join(log_dir, "params", "env.yaml"), env_cfg)
     dump_yaml(os.path.join(log_dir, "params", "agent.yaml"), agent_cfg)
-    dump_pickle(os.path.join(log_dir, "params", "env.pkl"), env_cfg)
-    dump_pickle(os.path.join(log_dir, "params", "agent.pkl"), agent_cfg)
 
     # run training
     runner.learn(num_learning_iterations=agent_cfg.max_iterations, init_at_random_ep_len=True)

scripts/reinforcement_learning/sb3/train.py

@@ -91,7 +91,7 @@ def cleanup_pbar(*args):
     multi_agent_to_single_agent,
 )
 from isaaclab.utils.dict import print_dict
-from isaaclab.utils.io import dump_pickle, dump_yaml
+from isaaclab.utils.io import dump_yaml
 
 from isaaclab_rl.sb3 import Sb3VecEnvWrapper, process_sb3_cfg
 
@@ -130,8 +130,6 @@ def main(env_cfg: ManagerBasedRLEnvCfg | DirectRLEnvCfg | DirectMARLEnvCfg, agen
     # dump the configuration into log-directory
     dump_yaml(os.path.join(log_dir, "params", "env.yaml"), env_cfg)
     dump_yaml(os.path.join(log_dir, "params", "agent.yaml"), agent_cfg)
-    dump_pickle(os.path.join(log_dir, "params", "env.pkl"), env_cfg)
-    dump_pickle(os.path.join(log_dir, "params", "agent.pkl"), agent_cfg)
 
     # save command used to run the script
     command = " ".join(sys.orig_argv)

scripts/reinforcement_learning/skrl/train.py

@@ -104,7 +104,7 @@
 )
 from isaaclab.utils.assets import retrieve_file_path
 from isaaclab.utils.dict import print_dict
-from isaaclab.utils.io import dump_pickle, dump_yaml
+from isaaclab.utils.io import dump_yaml
 
 from isaaclab_rl.skrl import SkrlVecEnvWrapper
 
@@ -168,8 +168,6 @@ def main(env_cfg: ManagerBasedRLEnvCfg | DirectRLEnvCfg | DirectMARLEnvCfg, agen
     # dump the configuration into log-directory
     dump_yaml(os.path.join(log_dir, "params", "env.yaml"), env_cfg)
     dump_yaml(os.path.join(log_dir, "params", "agent.yaml"), agent_cfg)
-    dump_pickle(os.path.join(log_dir, "params", "env.pkl"), env_cfg)
-    dump_pickle(os.path.join(log_dir, "params", "agent.pkl"), agent_cfg)
 
     # get checkpoint path (to resume training)
     resume_path = retrieve_file_path(args_cli.checkpoint) if args_cli.checkpoint else None

source/isaaclab/config/extension.toml

@@ -1,7 +1,7 @@
 [package]
 
 # Note: Semantic Versioning is used: https://semver.org/
-version = "0.46.5"
+version = "0.47.0"
 
 # Description
 title = "Isaac Lab framework for Robot Learning"

source/isaaclab/docs/CHANGELOG.rst

@@ -1,9 +1,22 @@
 Changelog
 ---------
 
+0.47.0 (2025-10-14)
+~~~~~~~~~~~~~~~~~~~
+
+Changed
+^^^^^^^
+
+* Removed pickle utilities for saving and loading configurations as pickle contains security vulnerabilities in its APIs.
+  Configurations can continue to be saved and loaded through yaml.
+
+
 0.46.5 (2025-10-14)
 ~~~~~~~~~~~~~~~~~~~
 
+Added
+^^^^^
+
 * Exposed parameter :attr:`~isaaclab.sim.spawners.PhysxCfg.solve_articulation_contact_last`
   to configure USD attribute ``physxscene:solveArticulationContactLast``. This parameter may
   help improve solver stability with grippers, which previously required reducing simulation time-steps.

source/isaaclab/isaaclab/utils/io/__init__.py

@@ -7,5 +7,4 @@
 Submodules for files IO operations.
 """
 
-from .pkl import dump_pickle, load_pickle
 from .yaml import dump_yaml, load_yaml