claudedwithlove
explore/francescofioredev-agentic-infra-claude-skill

agentic-infra-claude-skill

Cherished

A Claude Code skill that helps architects design, review, and validate agentic AI systems against 21 established patterns and a 62-item checklist. It provides deterministic linting of architecture decision records with 17 rules and structured guidance for teams building multi-agent systems.

·0··submitted April 19, 2026
View on GitHub
Clauded With Love Rating
7.7 / 10

This Claude Code skill provides structured guidance for designing and validating agentic AI systems using 21 established patterns and a comprehensive 62-item checklist. It includes a deterministic linter with 17 rules for architecture decision records and covers the full spectrum of multi-agent system concerns.

Code Quality6.5
Usefulness8.2
Claude Usage7.5
Documentation7.8
Originality8.7
Highlights
  • Comprehensive taxonomy of 21 agentic patterns with structured concern mapping
  • Deterministic linter with JSON output and proper exit codes for CI/CD integration
  • Evidence-based approach with PDF citations for every rule and recommendation
To Improve
  • Add actual code samples and test coverage for the lint_agentic_arch.py script
  • Provide concrete examples of the 62-item checklist in action with real ADR reviews
Language