run-experiment

Run experiments in isolated tmux windows with persistent logs.

11|2|Updated May 29, 2025
One-click install
npx skills add https://github.com/yulonglin/dotfiles --skill run-experiment-yulonglin
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: run-experiment
Source: https://github.com/yulonglin/dotfiles/tree/main/claude/local-marketplace/plugins/research-toolkit/skills/run-experiment
Command: npx skills add https://github.com/yulonglin/dotfiles --skill run-experiment-yulonglin

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Running large AI experiments often creates context pollution when run in the same workspace or session, making results hard to reproduce. This skill provides isolated, named tmux sessions and windows with persistent logs to ensure clean, repeatable runs.

Core Features & Use Cases

  • Isolated experimental windows: Each run gets its own tmux window within a shared session to prevent cross-talk.
  • Persistent logging: All outputs are captured to log files in tmp/ for later analysis.
  • Reproducible workflows: Recreates the exact session structure to facilitate repeatable experimentation.

Quick Start

Create an isolated tmux session and named window for your experiment, run your command, and tee output to a log file.

Frequently Asked Questions about run-experiment

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

FAQPage Schema
How do I run AI experiments in isolated tmux sessions to prevent context pollution?

Running AI experiments in isolated tmux sessions prevents context pollution by creating a dedicated named window within a shared session, ensuring clean and repeatable runs without cross-talk between different experiments.

What's the best way to keep persistent logs for reproducible experiment sessions across disconnects?

Persistent logs for reproducible experiment sessions are kept by teeing all command outputs to time-stamped log files stored in tmp/, preserving the data even across unexpected disconnects for later analysis.

Does tmux session management work for maintaining reproducible AI research workflows?

Tmux session management works for reproducible AI research workflows by recreating the exact session structure with isolated windows, which facilitates repeatable experimentation and prevents cross-contamination of results.

How do I log experiment outputs to a time-stamped file while running in a tmux window?

To log experiment outputs to a time-stamped file while in a tmux window, you run your command and tee the output directly to a designated log file in the tmp/ directory for later analysis.

When do I need isolated tmux windows for my experiment runs?

You need isolated tmux windows for experiment runs when executing large AI experiments that risk context pollution in a shared workspace, requiring clean and reproducible sessions to guarantee accurate results.