Development on Vibe Coding

Artifacts Builder

Mon, 01 Jan 0001 00:00:00 +0000

The Artifacts Builder skill covers the conventions for producing standalone, self-contained interactive outputs — inlined styles and scripts, responsive layout, no external dependencies that would break in a sandboxed viewer.

Webapp Testing

Mon, 01 Jan 0001 00:00:00 +0000

The Webapp Testing skill packages a verification workflow: launching a dev server, driving a browser through the golden path and edge cases, and checking console and network output for errors — rather than declaring a UI change complete after just reading the code.

MCP Builder

Mon, 01 Jan 0001 00:00:00 +0000

The MCP Builder skill packages the boilerplate and conventions for writing a new MCP server — the protocol handshake, tool and resource definitions, and error handling patterns that a well-behaved server needs.

Claude API

Mon, 01 Jan 0001 00:00:00 +0000

The Claude API skill is a language-aware reference for the Anthropic SDK and Messages API — model selection, adaptive thinking, streaming, prompt caching, token counting, and tool/MCP integration — with explicit guidance to verify API shapes against live docs rather than a model’s memorized training data.

Skill Creator

Mon, 01 Jan 0001 00:00:00 +0000

Skill Creator packages the full lifecycle of authoring an Agent Skill: drafting the skill, writing test prompts, running evaluations with variance analysis, reviewing results, and rewriting the skill based on feedback, then optimizing the skill’s description for more reliable triggering.

Superpowers

Mon, 01 Jan 0001 00:00:00 +0000

Superpowers packages a battle-tested library of 20+ Claude Code skills — test-driven development, systematic debugging, brainstorming techniques, git worktree management, and root-cause tracing — installable in one shot as a plugin marketplace.

Superpowers Lab

Mon, 01 Jan 0001 00:00:00 +0000

Superpowers Lab is the experimental proving ground for Superpowers — newer, less-settled techniques that graduate into the stable library once they hold up in real use.

Why a skill for this

Skill development benefits from the same staging discipline as software releases — trying new techniques in a lab environment before committing them to a widely-installed stable marketplace avoids regressions for everyone else.

Expo Skills

Mon, 01 Jan 0001 00:00:00 +0000

Expo Skills is the official skill set from the Expo team for building React Native apps with Expo, covering the framework’s specific conventions and tooling.

Why a skill for this

Being maintained by the Expo team itself means the skill tracks the framework’s actual conventions and release changes, rather than a third party’s best guess at Expo’s current best practices.

shadcn/ui Skill

Mon, 01 Jan 0001 00:00:00 +0000

The shadcn/ui skill gives Claude Code direct context on shadcn/ui’s components and enforces consistent usage patterns across a project.

Why a skill for this

shadcn/ui ships as copyable component source rather than an installed package, so consistency depends on the model actually knowing each component’s intended API. A skill scoped to shadcn/ui keeps generated UI code matching the library’s real patterns instead of guessing.

Get Shit Done

Mon, 01 Jan 0001 00:00:00 +0000

Get Shit Done is a lightweight meta-prompting and context-engineering system for Claude Code, built around spec-driven development rather than ad hoc prompting.

Why a skill for this

Spec-first workflows tend to produce more predictable output than freeform prompting, especially on larger features. Packaging that discipline as a skill makes it the default instead of something re-explained per task.

Skill Seekers

Mon, 01 Jan 0001 00:00:00 +0000

Skill Seekers automatically converts any documentation website into a ready-to-use Claude skill, turning scattered docs into structured, invokable reference material.

Why a skill for this

Most third-party libraries have documentation but no packaged skill. Automating the conversion means any well-documented tool can get a working skill without hand-authoring one from scratch.

Jules Delegate

Mon, 01 Jan 0001 00:00:00 +0000

Jules Delegate hands off coding tasks to Google’s Jules AI agent for async work — bug fixes, docs, and tests — directly on a GitHub repo.

Why a skill for this

Some tasks are better run asynchronously in the background than synchronously in the current session. This skill packages the hand-off so Claude Code can delegate rather than block on work another agent can do independently.

Agent Skills Open Standard

Mon, 01 Jan 0001 00:00:00 +0000

Agent Skills is the open SKILL.md standard for packaging reusable agent capabilities — originated by Anthropic and now adopted across Codex, Gemini CLI, Grok Build, Cursor, GitHub Copilot, and more.

Codex Skills System

Mon, 01 Jan 0001 00:00:00 +0000

OpenAI Codex’s built-in skills system discovers SKILL.md skills from ~/.codex/skills/ and installs new ones via $skill-installer, either from OpenAI’s curated catalog or any GitHub repo.

Why a skill for this

Having a first-party install path and curated catalog, rather than only community-assembled lists, is what makes a skills ecosystem practical to actually adopt inside Codex.

OpenAI Skills/Plugins Catalog

Mon, 01 Jan 0001 00:00:00 +0000

This is OpenAI’s official, curated repository of Codex skills and plugins — covering design tools, testing, security, observability, and one-command cloud deploys.

Why a skill for this

A first-party curated catalog is a stronger trust signal than an assembled community list — these skills are maintained by, or reviewed by, OpenAI itself rather than crowdsourced without a review bar.

Security Best Practices Skill

Mon, 01 Jan 0001 00:00:00 +0000

This curated Codex skill applies a secure-coding checklist during code generation and review, catching common vulnerability classes as code is written.

Why a skill for this

Security issues are cheaper to prevent during generation than to catch in a later review pass. Baking a checklist into the generation step catches classes of bugs before they’re ever committed.

Security Threat Model Skill

Mon, 01 Jan 0001 00:00:00 +0000

This curated Codex skill walks through producing a structured threat model for a system — assets, trust boundaries, and attack surfaces — rather than an ad hoc security discussion.

Why a skill for this

Threat modeling is only useful if it’s structured and complete; an unstructured conversation tends to miss entire categories of risk. A skill that enforces the methodology produces a threat model that’s actually useful to a security reviewer.

Skill Creator (OpenAI)

Mon, 01 Jan 0001 00:00:00 +0000

This system skill, auto-installed with Codex, scaffolds a new SKILL.md-compliant skill from a plain description of what it should do.

Why a skill for this

Getting a skill’s folder structure and frontmatter right by hand is easy to get subtly wrong. A scaffolding skill means authoring a new skill starts from a correct template instead of a blank file.

Google Antigravity Skills System

Mon, 01 Jan 0001 00:00:00 +0000

Google Antigravity’s skills system supports the same SKILL.md format as Claude Code and Codex, with both global (~/.gemini/config/skills/) and per-project (.agents/skills/) scopes.

Why a skill for this

Supporting the shared SKILL.md format means skills built for Antigravity aren’t locked to Google’s tooling — the same skill folder is portable to other compatible agents.

Gemini CLI Skills System

Mon, 01 Jan 0001 00:00:00 +0000

Gemini CLI natively supports the Agent Skills (SKILL.md) open standard, letting it load the same skill format used by Claude Code, Codex, and other compatible agents.

Why a skill for this

A shared skills format means a skill only has to be written once to work across Gemini CLI and other compatible agents, rather than maintaining separate versions per tool.

Grok Build Skills System

Mon, 01 Jan 0001 00:00:00 +0000

Grok Build’s skills system reads SKILL.md skills from .grok/skills/, ~/.grok/skills/, and plugin directories — and is directly Claude Code-compatible, reading .claude/skills and CLAUDE.md with zero extra configuration.

Why a skill for this

Direct compatibility with Claude Code’s own skill and config files means a project that’s already set up for Claude Code works in Grok Build without any migration step.