REPO PACKAGE
tw93 waza
COMMUNITY8skills
7ready to use
LABSCO SUMMARY
Each skill maps to one habit and one slash command: /think produces a decision-complete plan before code is written; /ui produces UI with a committed aesthetic direction rather than generic defaults; /check reviews a diff or gates a release; /hunt finds a bug's root cause before any fix is applied; /write rewrites prose in Chinese or English and strips AI-sounding phrasing; /learn runs a six-phase research pass over a source bundle; /read fetches and summarizes a URL or PDF; and /health audits Claude Code's, Codex's, or Pi's own configuration for drift. The README lays out a suggested chain for each of four common workflows — plan a feature, ship a fix, research and write, debug and verify — but every handoff between skills is a manual step the user triggers, not an automatic pipeline.
It's for someone already running Claude Code, Codex, Cursor, or a compatible harness who wants a fixed set of habits enforced consistently rather than reinventing them per session — and, going by the README's own framing, wants exactly these eight and no more. Waza's author built it smaller on purpose, contrasting it with heavier alternatives like Superpowers, on the theory that fewer, sharper skills beat a large library.
READ THE FULL ANALYSIS
What we can't verify. The README says the project is "refined through 300+ sessions across 7 projects" and that every documented gotcha traces to a real failure; that's the author's own account of Waza's history, not something we independently checked. Two of the eight skills need a local tool already on the machine rather than installing clean — our dependency check splits the eight six ready, two local_tool — but the descriptions don't say which two, so we can't tell you without opening each skill file.
ALSO IN THIS PACKAGE
PLUGINS
9 itemsPLUGIN
waza
Installs the full Waza toolkit. Registers all eight skills under the waza namespace, callable as /waza:think, /waza:check, /waza:hunt, /waza:ui, /waza:read, /waza:write, /waza:learn, and /waza:health.
PLUGIN
waza-check
Reviews diffs, PRs, release readiness, and publishing follow-through. Use when asked to review, triage issues or PRs, or ship.
PLUGIN
waza-health
Audits agent config, instruction drift, hooks or MCP, and AI maintainability. Use when Claude, Codex, or Pi setup looks wrong.
PLUGIN
waza-hunt
Finds root cause before any fix. Use when something errors, crashes, regresses, or used to work.
PLUGIN
waza-learn
Runs a six-phase research workflow from source bundle to publish-ready output. Use when researching an unfamiliar domain or compiling materials into one reference.
PLUGIN
waza-read
Fetches URLs and PDFs, then summarizes or returns clean Markdown. Use when asked to read, fetch, quote, cite, convert, or save a URL or PDF.
PLUGIN
waza-think
Turns rough ideas into approved, decision-complete plans before coding. Use when planning architecture, judging whether to build, or writing a handoff.
PLUGIN
waza-ui
Produces distinctive production UI and screenshot-grounded visual polish. Use when building or restyling pages, components, or typography.
PLUGIN
waza-write
Rewrites and polishes Chinese or English prose and product copy. Use when drafting, editing, localizing, or cutting AI tone.
WHAT'S INSIDE
8 showing · 8 totalReady to use
Nothing else to set up — install it and go.
8 of 8
SETUP
Showing all 8 skills
check
Reads the change you are about to ship and tells you what is wrong with it, and, only if you then say so, carries it the rest of the way to committed, pushed or released.
health
A check-up on the setup your AI coding assistant works under: the instruction files, permissions, hooks and connected servers it obeys, plus the corners of the codebase that have quietly become unsafe for it to touch.
hunt
A patch on the symptom only moves a bug somewhere else, so nothing gets changed here until the real cause is named in one sentence and backed by evidence anyone can re-run.
learn
Six steps from knowing nothing about a subject, or owning a pile of unread sources, to notes, a working mental model, or an article that is ready to publish.
read
Pulls the real text out of a web page or a PDF and hands it back in whatever form you asked for: a short summary, an exact quote with its source, clean Markdown, or a saved file.
think
Before any code gets written, an idea gets argued down to one recommended plan: what to build, which alternatives were rejected and why, and whether the thing is worth doing at all.
ui
Designs a screen that looks deliberate rather than like the default any prompt would produce, then keeps comparing the real rendered page against that intent until it holds up at every window size.
write
Edits writing back into something a person would plausibly have said, in Chinese or English, across drafts, docs, release notes and launch posts, without touching what it actually means.
No skill in this package matches that.
HOW TO GET IT
WHOLE PACKAGEInstall the whole package (8 skills)
npx skills add tw93/wazaSINGLE SKILLOr install a single skill
npx skills add tw93/waza --skill <name> --full-depthPick the skill name from the Skills tab — each entry there installs independently.
PLUGIN MARKETPLACEAdd the marketplace, then install the plugins you want
/plugin marketplace add tw93/waza/plugin install waza@waza/plugin install waza-check@waza/plugin install waza-health@wazaTyped inside the agent's own prompt, not in a terminal. The marketplace lists 9 plugins; three are shown. Every one installs the same way, with its own name before @waza.