Curated library of 16 public Ruby AI agent skills: 10 atomic skills (YARD docs, service objects, calculator pattern, API clients, DDD, bug triage, code review, skill routing), 5 process-discipline skills (TDD, refactoring, review, security, test planning), and 1 planning skill (TDD task generation). Zero agents — this is a foundational library consumed by framework-specific tiles like rails-agent-skills and hanakai-yaku.
95
96%
Does it follow best practices?
Impact
95%
1.05xAverage score across 16 eval scenarios
Passed
No known issues
Shared Ruby development skills and process-discipline knowledge for the AI skill ecosystem. This repository contains framework-agnostic foundations for TDD, refactoring, code review, security review, DDD, inline documentation, and common Ruby design patterns.
This repository is a core component of the multi-repo AI Skill Ecosystem:
ruby-core-skills (This repository): Contains 16 foundational Ruby programming, planning, and software engineering process-discipline skills. Contains zero agents.rails-agent-skills: Curated library of Rails-specific skills + 9 specialized agents that compose core processes.hanakai-yaku: Curated library of Hanami-specific skills + 10 specialized agents.agnostic-planning-skills: Generic project management, planning, and task breakdown skills + 4 agents.agent-mcp-runtime: The Rust-based CLI runtime that acts as the composition and resolution engine.flowchart BT
subgraph core["ruby-core-skills"]
direction LR
process["process skills"]
atomic["atomic skills"]
end
rails["rails-agent-skills<br/>(skills+agents)"]
hanakai["hanakai-yaku<br/>(skills+agents)"]
agnostic["agnostic-planning-skills<br/>(skills+agents)"]
rails --> core
hanakai --> core
agnostic --> coreFramework repos depend on core skills. ruby-core-skills does not know about any downstream frameworks.
Supported agent environments
| Skill | Category | Description |
|---|---|---|
| define-domain-language | DDD | Extracting ubiquitous language or glossary definitions. |
| review-domain-boundaries | DDD | Auditing context boundaries and language leakage. |
| model-domain | DDD | Tactical DDD design (aggregates, entities, value objects, domain services). |
| write-yard-docs | Documentation | Writing or reviewing inline YARD documentation for public Ruby APIs. |
| create-service-object | Patterns | Creating a service object (PORO .call pattern). |
| implement-calculator-pattern | Patterns | Implementing polymorphic variant-based calculators (Strategy + Factory). |
| integrate-api-client | Patterns | Designing HTTP integrations (layered client/fetcher/builder pattern). |
| triage-bug | Testing | Investigating a bug, reproducing via failing test, and creating a repair plan. |
| respond-to-review | Code Quality | Receiving code review feedback and addressing comments. |
| skill-router | Orchestration | Triaging and decomposing complex Ruby requests into ordered sub-tasks. |
| generate-tdd-tasks | Planning | Breaking features into TDD quadruplet task lists with docs and review tasks. |
| tdd-process | Process | General engineering loop: Red-Green-Refactor process gates and checkpoints. |
| refactor-process | Process | Safely refactoring code while preserving behavior under characterization tests. |
| review-process | Process | Reviewing changesets (severity taxonomies, structured findings, re-review). |
| security-review-process | Process | Reviewing code for general Ruby security flaws (secrets, injections). |
| test-planning-process | Process | Choosing test boundaries (unit vs integration) and test scenarios. |
Core skills are consumed by AI agents via the agent-mcp-runtime CLI tool.
During development, you can point the runtime to this local directory using the --registry flag:
agent-mcp-runtime --registry ./path/to/ruby-core-skills --task "Review this code"In production, the runtime uses the central registry.json manifest to resolve dependencies automatically. When running a framework pack, the runtime loads the framework's own skills first, falling back to core skills for any general Ruby concerns:
# Resolves Rails agents/skills first, then falls back to core skills
agent-mcp-runtime --pack rails --task "Add full_name to User model"This project is licensed under the MIT License. See LICENSE for details.