Remote Observable Process
DevOps & Securityremote-observable-processRemote Observable Process is an Agent Skill for devops & security published in the Mesh-LLM/mesh-llm repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.
Use this skill when starting, supervising, debugging, holding open, or stopping any remote process over SSH that needs an operator-like interactive environment, a TTY, login-shell startup files, long-running observation, logs, readiness checks, or later inspection.
How to use this skill
Bring this guide into your coding agent with a prompt tailored to the tool you use.
- Open your project in Codex.
- Copy and paste the prompt below.
- Review files and risks before approving any action.
Inspect the Agent Skill at https://github.com/Mesh-LLM/mesh-llm/blob/HEAD/.agents/skills/remote-observable-process/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/remote-observable-process, but wait for my approval before writing files or running scripts.
Copying this prompt does not install or run the skill. Review third-party files before use.
remote-observable-process
Use this skill before starting a non-trivial remote process over SSH when the process needs to behave like it was launched by an operator in a terminal, be observed after launch, expose readiness, keep running beyond one command, or be stopped cleanly later.
Rule
For environment-sensitive processes, prefer SSH with a TTY and an interactive login shell:
ssh -tt host '/bin/zsh -ilc '\''COMMAND'\'''
If the remote host does not use zsh, adapt the shell while preserving the same properties: allocate a TTY, use a login/interactive shell, and keep the session foreground for first repro/debug runs.
Avoid detached first attempts such as:
ssh host "nohup COMMAND > /tmp/process.log 2>&1 &"
ssh host "COMMAND > /tmp/process.log 2>&1 &"
Those shapes hide lifecycle and can behave differently from a real remote session.
Stage Runtime Debugging
For network-sensitive server chains, model stage servers, GPU/Metal workloads, or bind/connect debugging, first prove the process in a held foreground TTY:
ssh -tt host '/bin/zsh -ilc '\''COMMAND 2>&1 | tee /tmp/COMMAND.log'\'''
Keep the SSH session open while sending traffic. Use tmux or screen only
after validating that they preserve the same listener, downstream connection,
and request behavior on that host.