Synchronize DeepResearch experiment code to remote HPC and cloud sandboxes.
npx skills add https://github.com/dongzhuoyao/deepresearch --skill sue-synccode
Please help me install this Agent Skill. Skill: sue-synccode Source: https://github.com/dongzhuoyao/deepresearch/tree/main/.codex/skills/sue-synccode Command: npx skills add https://github.com/dongzhuoyao/deepresearch --skill sue-synccode
Manually syncing experiment code to remote sandboxes is error-prone, risks overwriting valuable run artifacts, and wastes time debugging SSH access or git authentication issues. This Skill eliminates those risks by enforcing strict path contracts and automated sync workflows.
Use the sue-synccode skill to sync your latest DeepResearch experiment code to the selected remote sandbox before running dryrun or fullrun validation.
High-intent search queries and answers about installing and using this skill.
To sync experiment code without overwriting run artifacts, use automated synchronization that enforces strict workspace identity contracts and applies inode-safe file exclusion rules. This prevents path mismatches and preserves previous checkpoint and result files during the transfer.
Yes, rsync works for pushing ML experiment code if SSH fails. The sync workflow supports both git and rsync or scp transfer methods, automatically falling back to rsync or scp when sandbox GitHub SSH authentication fails so code delivery is never blocked.
The best way to save inodes during remote HPC transfers is to apply automatic inode-safe exclusion rules. This process excludes virtual environments, caches, and runtime artifacts from the synchronization, preventing unnecessary file count bloat on the sandbox.
Yes, you can use this code synchronization method for pre-dryrun, pre-fullrun, retry, and remote validation workflows. It supports all project-supported sandboxes including LUMI, Snellius, NM5, Brev, RunPod, and AutoDL for reliable experiment code delivery.
You need strict path governance when syncing launcher scripts to prevent path mismatches that could break experiment runs or corrupt outputs. Enforcing DeepResearch root and workspace identity contracts ensures code lands in the correct directory structure safely.
Your checkpoint files remain safe when you push latest code changes. The synchronization process automatically excludes runtime artifacts and preserves existing experiment outputs, ensuring that only source code and launcher scripts are updated on the remote sandbox.
JuanMS20Configure GitHub authentication for Hermes Agent using git, gh CLI, or SSH credentials.
llv22Launch and manage ML experiments on local or remote GPU servers.
PBH-BTNArchive completed experimental changes with validated artifacts and tasks.
AirmomoManage SSH keys, tunnels, and file transfers with OpenSSH tools.
