remote-run

Execute commands, builds, and test suites headlessly on a remote machine over SSH.

Updated Feb 6, 2026
One-click install
npx skills add https://github.com/kellymears/agents --skill remote-run-kellymears
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: remote-run
Source: https://github.com/kellymears/agents/tree/main/plugins/remote-run/skills/remote-run
Command: npx skills add https://github.com/kellymears/agents --skill remote-run-kellymears

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Heavy workloads like full test suites, Docker stacks, asset builds, and local LLM inference can overwhelm a local machine. This Skill offloads that work to a configured remote dev box over SSH, running everything headlessly and reporting output back, so your local environment stays responsive. ## Core Features & Use Cases - Headless remote execution: Runs commands with no TTY, no interactive prompts, and non-interactive environment variables, capturing output and reporting which host executed the work. - Config-driven host management: Resolves hosts from a per-user or per-project JSON config with SSH fallbacks, path mappings, and host-specific notes, plus a persistent per-host memory file for lessons learned across sessions. - Smart code sync and job control: Chooses between git checkout, rsync mirroring, or ad-hoc script transfer, detaches long-running jobs with nohup and PID files, supports port forwarding, and copies artifacts back when needed. - Use Case: Ask to run your full test suite on your build box; the Skill verifies SSH connectivity, syncs your working tree, launches the suite detached, polls the log for progress, and reports the results. ## Quick Start Ask the agent to run the full test suite on your remote dev box and report the results back. First-time users can run /remote-run:setup for guided configuration of their SSH target and remote paths.

Frequently Asked Questions about remote-run

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run commands on a remote machine over SSH without interactive prompts?▼

Use ssh with the -T flag to disable pseudo-terminal allocation and BatchMode=yes to prevent password or hostkey prompts. Export non-interactive variables like CI=1 and GIT_TERMINAL_PROMPT=0, and prefer each tool's non-interactive flags such as --yes or --no-input.

How do I run long-running jobs on a remote server and check progress?▼

Detach the job with nohup, redirect output to a log file, and save the PID: nohup command > job.log 2>&1 & echo $! > job.pid. Then poll with tail on the log and check liveness with kill -0 on the stored PID.

How do I sync local code changes to a remote dev box?▼

If the remote repo exists and your tree is clean, fetch and check out the exact commit over SSH. For dirty trees or unpushed commits, use rsync -az --delete with a .gitignore filter so build artifacts and node_modules stay local.

Can I access a remote dev server or database from my local machine?▼

Yes, start the service detached on the remote host, then forward the port with ssh -f -N -L local-port:localhost:remote-port host. The service becomes reachable at the local URL while the tunnel stays open.

What happens if the remote host is unreachable over SSH?▼

The Skill tries each configured ssh_fallbacks target in order, such as a Tailscale alias for off-LAN access. If all targets fail, it reports which were tried and stops rather than silently running the work locally.

Where is the remote host configuration stored?▼

Configuration resolves in priority order: the REMOTE_RUN_CONFIG environment variable, a project-level .claude/remote-run.json, then ~/.claude/remote-run/config.json. The REMOTE_RUN_HOST variable can override the host for a single session.