compare-repos

Synthesizes cross-repo structural comparisons from per-repo analysis documents.

Updated Apr 19, 2026
One-click install
npx skills add https://github.com/nickgogan/improvement-loop --skill compare-repos-nickgogan
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: compare-repos
Source: https://github.com/nickgogan/improvement-loop/tree/main/.claude/skills/compare-repos
Command: npx skills add https://github.com/nickgogan/improvement-loop --skill compare-repos-nickgogan

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? When you track multiple watched libraries, understanding how they converge, diverge, or contradict each other requires manually reading every per-repo analysis. This Skill reads the existing *-analysis.md files and produces a citation-grounded comparison matrix, pattern clusters, and optional adoption recommendations. ## Core Features & Use Cases - Structural Comparison Matrix: Builds per-dimension tables (structural scale, context files, workflow topology, governance, cross-agent protocols, research dimension mapping) with citations to each repo's analysis doc. - Pattern Cluster Detection: Classifies patterns as shared conventions (3+ repos), unique innovations (1 repo), or contradictory approaches, all traceable to source sections. - Builder-Mode Recommendations: With --focus, adds a recommendation section covering trade-offs, pitfalls, and follow-up guides for a specific adoption question. - Use Case: After analyzing five agentic-coding libraries with /repo-analyzer, ask which governance model to adopt; the Skill compares all five and recommends patterns with cited trade-offs. ## Quick Start Ask the assistant to compare the watched libraries on governance models and recommend which approach to adopt for your project.

Frequently Asked Questions about compare-repos

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

FAQPage Schema
How do I compare multiple watched libraries across structural dimensions?▼

Run the comparison after each library has a per-repo analysis doc. The Skill reads all *-analysis.md files, builds dimension-by-dimension matrices, and clusters shared, unique, and contradictory patterns with citations to source sections.

How to get adoption recommendations from a repo comparison?▼

Pass a focus question with the --focus argument. The report then adds a Builder-mode recommendation section covering which repo's pattern fits your situation, relevant trade-offs, pitfalls from KB findings, and follow-up guides.

What happens if a library has no analysis document yet?▼

The comparison stops and reports the missing input, asking you to run the per-repo analyzer first. It requires at least two analyzed repos and never re-derives structural facts from raw repositories itself.

Does the comparison report get written to disk automatically?▼

No. Reports are presented inline by default. The --write flag only enables the write option; the actual write to cross-repo-comparison.md requires explicit approval, and prior versions are preserved in a Version Log.

When should I not use cross-repo comparison?▼

Do not use it for analyzing a single repo, promoting findings into the knowledge base, or live web research. Those belong to the per-repo analyzer, the findings-promotion skill, and the web research comparison workflow respectively.