igmarin/rails-agent-skills

Curated library of 28 public AI agent skills for Ruby on Rails development. Organized by category: testing, code-quality, engines, infrastructure, api, and context. Covers code review, architecture, security, testing (RSpec), engines, Hotwire, and TDD automation. Shared Ruby skills (YARD docs, DDD, service objects) have moved to ruby-core-skills. Repository agents remain documented in GitHub but are intentionally excluded from the Tessl tile.

1.78x

Quality

95%

Does it follow best practices?

Impact

93%

1.78x

Average score across 28 eval scenarios

Securityby

Passed

No known issues

Rails Agent Skills Documentation

Name: igmarin/rails-agent-skills
Rating: 93.21 (1 reviews)
Author: igmarin

This documentation is the public map for the Rails Agent Skills library: 28 public Rails-specific skills, 9 callable agents, 15 core Ruby skills (from igmarin/ruby-core-skills), an MCP distribution, and the evaluation process used to keep the skills useful.

For the high-level project value proposition, start with the root README.

Quick Start Decision Map

If you are...	Use this
New to the project	`load-context` and Discovery & Context
Ready to build a feature	Development Guide
Reviewing code	Review Guide
Not sure where to start	`skill-router` (from ruby-core-skills)
Installing the library	Root README
Validating skill quality	Eval provenance and Skill optimization

Note: This repository depends on igmarin/ruby-core-skills for foundational Ruby skills. Install both repositories for complete functionality.

Architecture: Skills vs. Agents

The repository uses a hybrid model to keep agent context focused:

Skills are atomic expert instructions such as code-review, plan-tests, or create-service-object. Agents load them on demand through an installed skill host.
Agents are orchestrated multi-step processes that chain multiple skills together into a complete development loop, such as TDD feature work, code review, setup, quality checks, or engine development.

Master Stage Index

Stage	Guide	Description	Primary skills
Discovery	Discovery & Context	Understand codebase and onboarding context	`load-context`, `setup-environment`
Setup	Setup & Configuration	CI/CD and infrastructure references	`setup-environment`
Development	Development	TDD and implementation	`plan-tests`, `write-tests`, `triage-bug` (from core)
Quality	Code Quality	Conventions, refactoring, and docs	Local: `apply-code-conventions`, `refactor-code`. Core: `write-yard-docs`, `refactor-process`, `review-process`
Review	Review & Validation	Review, security, and architecture	`code-review`, `security-check`, `review-architecture`, `respond-to-review` (from core)
Engines	Engine Development	Building and releasing Rails engines	`create-engine`, `release-engine`

Core Principles

Tests Gate Implementation

Implementation code waits until a test exists, has run, and fails for the expected reason. This is the central quality rule across code-producing skills.

Skill and Agent Chaining

Skills are building blocks. Agents define the sequence. Follow skill-router or the integration matrix when the next step is unclear.

Evaluation Ownership

Tessl validates publishable skills from tile.json using tessl-evals/. The upcoming ruby-skill-bench gem will use personal-evals/ for full-context skill and agent validation. Root evals/ is generated Tessl staging output and should not be committed.

Reference & Authoring

Skill Catalog - Complete list of 28 public Rails skills and 9 callable agents (plus 15 core skills from ruby-core-skills).
Integration Matrix - How skills connect.
Agent Guide - Narrative guide for daily use.
Architecture - Repository layout and SKILL.md conventions.
Eval Provenance - Canonical eval ownership policy.
Skill Template - Template for creating new skills.