The Agent Skills Directory

[SAFE]: The skill contains technical documentation and configuration snippets for distributed training resiliency. No indicators of prompt injection, data exfiltration, or obfuscation were found.
[COMMAND_EXECUTION]: Provides standard command-line instructions for launching distributed training jobs and running verification tests.
Uses ft_launcher for fault-tolerant job execution.
Includes commands to run example scripts located within the repository (e.g., ./examples/training_features/...).
Uses uv and pytest for executing verification tests and straggler detection examples.
[EXTERNAL_DOWNLOADS]: Identifies nvidia-resiliency-ext as a required dependency. This is a legitimate library for the training resiliency features described in the skill.

nemo-mbridge-resiliency