Explore Run

When your deep learning researcher says "try a few quick variants" or "sweep this on a subset," this is the skill that plans and documents those exploratory runs without contaminating your trusted baseline. It ranks candidates by cost, success rate, and expected gain, then writes structured reports in explore_outputs/ that explain what you tried and why the comparisons might not be apples to apples. The boundaries are clear: it's for explicit exploration only, not production training or implicit experiments. Honestly, the real value is in the discipline it enforces around labeling exploratory evidence as exploratory, which matters more than you'd think when you're six months into a project trying to remember which runs were serious.

skills.sh installs: 352.7K
GitHub stars: 486
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add lllllllama/rigorpilot-skills --skill explore-run --agent claude-code

Installs into .claude/skills of the current project.

Files

explore-run

Use this as the Rigor Improve / Rigor Explore run leaf skill. The installed slug remains explore-run for compatibility.

Use the shared operating principles in ../ai-research-reproduction/references/agent-operating-principles.md; this skill should guide candidate run planning while preserving model judgment about the active repo.

When to apply

When the researcher explicitly authorizes exploratory runs.
When the task is a small-subset validation, short-cycle training probe, batch sweep, idle-GPU search, or quick transfer-learning trial.
When the output should rank candidate runs rather than certify trusted success.