agentic-infra-claude-skill
♥Cherished
A Claude Code skill that provides structured guidance for building agentic AI systems through design assistance, architecture review, and automated linting. It covers 21 established patterns for agent workflows, communication, safety, and adaptation—enabling developers to audit existing designs against a 62-item checklist and validate ADR documents using deterministic rules.
Clauded With Love Rating
7.8 / 10
A Claude Code skill that provides structured guidance for building agentic AI systems through 21 design patterns, architecture review checklists, and automated linting. The project addresses a real gap in agentic system design by offering systematic methodology with evidence-based recommendations citing specific page numbers from source materials.
Code Quality6.0
Usefulness8.5
Claude Usage7.5
Documentation8.0
Originality9.0
Highlights
- ✓Comprehensive taxonomy of 21 agentic patterns covering the full spectrum from workflow to safety concerns
- ✓Evidence-based approach with all recommendations citing page numbers from source PDFs
- ✓Practical tooling including deterministic linter with 17 rules and structured 62-item review checklist
To Improve
- →Add actual code examples or test coverage to validate the lint_agentic_arch.py implementation quality
- →Include concrete case studies or examples showing before/after ADR improvements using the patterns