explore-code

Isolate exploratory code changes on dedicated branches with traceable mappings.

515|11|Updated Mar 30, 2026
One-click install
npx skills add https://github.com/lllllllama/ai-research-workflow-skills --skill explore-code
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: explore-code
Source: https://github.com/lllllllama/ai-research-workflow-skills/tree/main/skills/explore-code
Command: npx skills add https://github.com/lllllllama/ai-research-workflow-skills --skill explore-code

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

Isolate exploratory code changes on dedicated branches to protect the trusted baseline while providing auditable records of what was proposed, tested, and reasoned about during research workflows.

Core Features & Use Cases

  • Isolates exploratory edits on separate branches or worktrees to prevent contamination of the trusted baseline.
  • Provides a structured plan and traceable mappings from candidate edits to target components, with notes on scope and rationale.
  • Outputs a compact bundle of summaries (CHANGESET.md, TOP_RUNS.md, status.json) suitable for review, auditing, and reproducibility.

Quick Start

Run the explore-code script on an isolated branch to generate explore_outputs with a detailed change plan.

Frequently Asked Questions about explore-code

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

FAQPage Schema
How do I isolate exploratory code changes to protect my trusted baseline?

Isolate exploratory code changes by operating on dedicated branches or worktrees, which prevents contamination of the trusted baseline while maintaining traceable mappings from candidate edits to target components.

What is the best way to maintain audit-ready records for research code modifications?

Maintain audit-ready records by generating a compact bundle of summaries including CHANGESET.md, TOP_RUNS.md, and status.json, which document proposed edits, testing results, and reasoning for reproducibility.

How do I ensure exploratory edits remain candidate-only until approved?

Exploratory edits remain candidate-only through enforced guardrails that require explicit authorizations for exploration, ensuring proposed changes cannot merge into the baseline until formal approval is granted.

Can I use isolated worktrees to plan and track code exploration in research repositories?

Yes, you can use isolated worktrees to run exploration scripts that generate detailed change plans and structured outputs, providing traceable mappings and rationale notes for candidate edits.

What outputs do I get from isolating exploratory edits on a separate branch?

You receive a compact bundle of explore_outputs containing CHANGESET.md, TOP_RUNS.md, and status.json, which are suitable for review, auditing, and ensuring the reproducibility of research workflows.