JustineDevs Skills Collection

A collection of AI agent skills for project development and execution across Codex, Claude, Composer, and similar agent runtimes.

Installation Collection

Install	Skills (URL)	Category	Codex	Purpose	Demo Prompt
npx skills add https://github.com/justinedevs/collection --skill blueprintkit	justinedevs/collection/blueprintkit	Planning	Yes	Complete project planning and execution framework: 14 planning sections plus 9 specialized execution capabilities, from initial vision through deployment and continuous improvement.	Create a project plan using BlueprintKit: scope, timeline, risk section, and recommended execution workflow.
npx skills add https://github.com/justinedevs/collection --skill erc8004-agent-creator	justinedevs/collection/erc8004-agent-creator	Agents	Yes	Usage and installation for create-8004-agent: how to create ERC-8004/8004 agents, wizard options, chains (EVM and Solana), and post-generation steps. Optional script to switch generated project to Anthropic Claude.	How do I create an ERC-8004 agent, what wizard options matter, and how do I register and run it after generation?
npx skills add https://github.com/justinedevs/collection --skill core-engineering	justinedevs/collection/core-engineering	Engineering	Yes	Core Engineering: OOP four pillars (Abstraction, Encapsulation, Inheritance, Polymorphism), SOLID, composition over inheritance, DRY/KISS/YAGNI/Law of Demeter, and association/aggregation/composition with example code.	Explain SOLID with code examples, then refactor this class to follow Single Responsibility.
npx skills add https://github.com/justinedevs/collection --skill context-engineering	justinedevs/collection/context-engineering	Prompting	Yes	Context Engineering: Skeleton-of-Thought, token-efficient diff updates, XML tagging, symbolic rule variables, and negative constraints, plus guidance on LLMBundle-style context packaging.	Show me a Skeleton-of-Thought outline for this repo and give me SEARCH/REPLACE diffs only for the changed function.
npx skills add https://github.com/justinedevs/collection --skill co-star-ui	justinedevs/collection/co-star-ui	UI	Yes	CO STAR UI: structured CO-STAR prompting for building/integrating React UI (shadcn, Magic UI, Aceternity, Newsprint) or creating custom CO-STAR prompts; discovery first, then plan and implementation.	Add a hero section using CO STAR and help me integrate the Newsprint design system into this React app.
npx skills add https://github.com/justinedevs/collection --skill agent-hub	justinedevs/collection/agent-hub	Orchestration	Yes	agent-hub: orchestration canon for the monorepo (architecture layers 00–10, indexes, diagrams, templates, command specs, agent operating rules). Machine entry: agent-hub/SKILL.md.	Use agent-hub to decide the correct orchestration path, architecture layer, and command flow for this repo task.
npx skills add https://github.com/justinedevs/collection --skill startup	justinedevs/collection/startup	Startup	Yes	Curated startup operating system bundle with startup-manager for request routing across founder thesis, validation, MVP scoping, fundraising, launch, growth, and founder ops.	Pick the one practical startup module I should use for this request, then help me validate the idea or scope the next blocking step.

Each skill is installed independently - you only download what you need.

Category groups similar skills. Codex = Yes means the published package includes a runtime skill folder that can be installed into ~/.codex/skills/.

Codex

Codex can use these published skills too. After downloading or copying a skill package, place the runtime skill folder into ~/.codex/skills/<skill-name>/.

Example for the startup bundle:

mkdir -p ~/.codex/skills/startup
cp -R startup/skills/startup/. ~/.codex/skills/startup/

The same idea applies to other agent runtimes: install the published package, then place the runtime-specific skill folder where that host discovers skills.

How Monorepo Installation Works

This repository uses a monorepo structure where multiple skills can coexist. When you install a skill:

• Selective Download: Only the specific skill directory is downloaded
• Independent Installation: Each skill can be installed/updated/removed separately
• Shared Resources: Skills can share common documentation or utilities at the collection level

The Installation Collection table above includes a demo prompt for each skill so users can quickly understand the real use case and expected workflow.

Result URL: https://skills.sh/justinedevs/collection/

Repository Structure

This is a monorepo containing multiple skills. Each skill can be installed independently. For the full file tree and directory layout, see STRUCTURE.md.

Contributing

See CONTRIBUTING.md for how to contribute to any skill, keep collection docs (README, STRUCTURE) accurate, and use optional workspace tooling (version/, .cursor/) in line with this repo.

Adding New Skills

To add a new skill to this collection:

1. Create a new directory (e.g., playbook/)
2. Add SKILL.md with the skill definition
3. Add README.md with skill documentation
4. Add skill content in .claude/skills/[skill-name]/ if needed
5. Update this README.md: add a row to the Installation Collection table
6. Update STRUCTURE.md and follow CONTRIBUTING.md

License

MIT License - See LICENSE file for details.