Overview
A collection of agent skills for AI coding agents, distributed via the skills CLI ecosystem. Skills are reusable instruction packages that teach agents how to handle specific types of work — with phases, guardrails, and success criteria. The collection currently includes an architecture decision framework, with more domains planned.
Why I Built This
AI agents are only as good as the methodology they follow. I wanted a way to package engineering discipline patterns — like structured decision-making — into sharable, installable instructions that any agent can load on demand. The skills CLI makes distribution trivial: one command and the skill is available for any session.
Key Features
Domain-Specific Methodology
Each skill encodes a complete workflow for a specific task type — with phases, guardrails, and success criteria — so agents follow a repeatable process rather than improvising.
Build Notes
General Agent-Skills Format
The project follows the general agent-skills specification — markdown content with YAML frontmatter, registered via a plugin manifest.
Reference Document Pattern
Heavy reference data (scoring profiles, templates, lookup tables) lives in companion files that agents fetch on demand, keeping the main skill focused on the workflow.