This repository ships Claude Code marketplace plugins. The plugins share a common focus on structured planning, reliable execution, and thoughtful polish in agent-assisted development.
Install the marketplace
/plugin marketplace add WeZZard/skillsInstall the plugins
/plugin install amplify@wezzard-skills # Amplify
/plugin install zelda-sounds@wezzard-skills # Zelda-Sounds
/plugin install skill-kit@wezzard-skills # Skill-Kit
/plugin install attune@wezzard-skills # Attune
/plugin install dispatch@wezzard-skills # Dispatch
/plugin install cupertino-taste@wezzard-skills # Cupertino-Taste
/plugin install workflows@wezzard-skills # Workflows
/plugin install x-articles@wezzard-skills # X-Articles
/plugin install additive@wezzard-skills # AdditiveDevelopment workflow skills for planning and execution.
| Skill | Description |
|---|---|
brainstorming |
Explore ideas, approaches, and requirements before implementation |
write-plan |
Create and update plan files with structured templates |
execute-plan |
Execute a plan file step by step |
same-page |
Explain a previous message with adaptive layout, evidence, and confidence |
be-thorough |
Investigate deeply before concluding when debugging or reviewing uncertain claims |
divide-and-conquer |
Break large jobs into parallel subagent-driven DAG workflows |
Zelda BotW and TotK sound cues for Claude Code lifecycle events, with a GUI configurator. Run /zelda-sounds:configure-zelda-sounds to assign sounds to hook events.
Tools for auditing and improving Claude Code skill definitions. Run /skill-kit:skill-lint --agent <agent> <path> to lint skill files for structural and schema issues.
George Orwell's rules on writing well, injected at session start to govern every reply and every document the agent writes, plus two user-invoked skills for the questions a discussion turns on: explore drives the design conversation and brings the user only what only they can settle; experiment settles by blind comparison.
Delegate tasks to external agent CLIs through one router and one brief contract — probed facts before use, write isolation through worktrees, a multi-model audit panel, and concurrent image generation with failure memory.
Distilled by WeZZard in China, Assembled from WWDC.
Run shared WeZZard workflows from local coding agents without copying workflow infrastructure into each repository.
Publish Markdown files as X Articles via XMCP and the official X API v2 Articles endpoints — deterministic conversion, image upload, draft-then-publish with a human checkpoint.
Additive agentic engineering — make each increment leave the system more verifiable than it found it. Design documents settle before code is written, architecture decisions live permanently in docs/adrs/ and are superseded rather than edited, and the permanent corpus is checked against the implementation rather than against prose style. Also carries the standing communication guidelines, selected by platform, model, and reply language.
MIT — see LICENSE.