JZKK720
- 44 skills
- 0 followers
- 4 repo stars
- yesterday last updated
- ▌ Impeccable Quieter · jzkk720Reduce visual intensity in a design that is too loud, aggressive, or overstimulating without losing personality or collapsing into generic. Use when a surface demands attention it has not earned, or reads exhausting to use. Covers desaturating and softening palettes, reducing visual weight, cutting decorative motion and effects, and refinement through restraint.
- ▌ Markitdown Testing · jzkk720Write and run tests for MarkItDown converters. Use when: adding tests for a new converter, writing parametrized test vectors, running pytest with hatch, understanding FileTestVector pattern, or debugging test failures. Covers test file layout, FileTestVector dataclass, parametrized test patterns, remote test skipping, and coverage.
- ▌ Impeccable Colorize · jzkk720Introduce color into an interface as hierarchy, meaning, and atmosphere rather than decoration. Use when a surface is under-colored, when color needs to encode action, selection, status, or wayfinding, or when a palette needs to carry atmosphere. Preserves confirmed brand and semantic conventions instead of replacing the visual world.
- ▌ Impeccable Optimize · jzkk720Find and fix the actual performance bottleneck in an interface, then measure the result. Use when an interface is measurably slow, or when performance is a stated requirement. Covers Core Web Vitals (LCP, INP, CLS), load time, bundle size, render cost, and the discipline of not optimizing what is not slow.
- ▌ Github Repo Metadata · jzkk720Use when updating a GitHub repository's description, homepage, topics, or other metadata via the API — especially when the gh CLI is unavailable and no GITHUB_TOKEN is set, so authentication must come from the local git credential store. Also covers the read-after-write caching trap that makes a successful PATCH look like a silent no-op.
- ▌ Impeccable Overdrive · jzkk720Push an interface past conventional limits using the full power of the platform - a table handling a million rows, a dialog morphing from its trigger, streaming form validation, cinematic page transitions. Use when a surface should feel technically extraordinary. Requires judging what extraordinary means for the specific project, since the same effect that impresses on a portfolio embarrasses on a settings page.
- ▌ Evidence Graph Review · jzkk720Evidence Graph Review — audit evidence-to-finding-to-path traceability, confidence grading, scope gating, and stop/replan conditions when authorization or evidence quality is insufficient. Methodology-only; no scripts, no tool execution.
- ▌ Impeccable Craft Floor · jzkk720The non-negotiable quality floor and absolute bans to verify immediately before shipping any UI. Use as a final pre-ship checklist on a built result, not as an intention. Covers contrast ratios, real shadow depth, focus visibility, motion respect, and the mechanical defects that design review tends to miss.
- ▌ Impeccable Adapt Native · jzkk720Recontextualize an existing native design (iOS / Android / adaptive) for a different device class, orientation, platform, or origin. Use when adapting a native app, where the trap is treating adaptation as scaling. Covers size classes, window size classes, posture and hinge handling on foldables, and translating idioms rather than transplanting them between iOS and Android.
- ▌ Impeccable Audit Native · jzkk720Run systematic technical quality checks on a native app (iOS / Android / adaptive) and generate a scored report without fixing anything. Use when auditing native implementation quality from source (SwiftUI, UIKit, Compose, React Native, Flutter). Scores five dimensions 0-4 against platform conventions, including VoiceOver and TalkBack accessibility.
- ▌ Ops Authorization Methodology · jzkk720Authorization and scope methodology for security/penetration/red-team tasks. Read-only checklist: confirm scope.md, network_profile mode, in-scope/out-of-scope assets, deliverables, and stop/replan conditions. No scripts, no toolchain coupling.
- ▌ Wigolo · jzkk720Local-first web intelligence — search, fetch, crawl, extract, research, and monitor web content. Use when gathering information from the web, fetching pages as markdown, crawling sites, extracting structured data, researching topics, or monitoring pages for changes. MCP server via `wigolo` (stdio) or `wigolo serve` (HTTP).
- ▌ Jev Harness · jzkk720 bundleBuild, evaluate, and optionally evolve task-specific code and Jev decision pipelines with JevHarness. Use when a user wants a reusable Jev agent, a new task adapter, or trajectory-and-reward optimization; begin by clarifying the task and evaluation contract. Not for ordinary coding without a Jev workflow.
- ▌ Watch Skill · jzkk720Video intelligence skill — watch, remember, and verify video content. Use when analyzing video files, extracting frames, searching video moments, transcribing audio, detecting objects or scenes, or building video-aware agent workflows. MCP server via `watch-skill serve`.
- ▌ Impeccable IOS · jzkk720iOS and iPadOS platform conventions for native apps built with SwiftUI, UIKit, React Native, Expo, or Flutter. Use when building or reviewing iOS UI. Covers Human Interface Guidelines conformance as the governing rulebook, expressing brand through the layer the platform leaves open, and the iOS slop test for controls that read as ported from a website.
- ▌ Impeccable Adapt · jzkk720Recontextualize an existing web design for a different screen size, device, platform, or use case. Use when adapting rather than extending a design - the trap is treating adaptation as scaling, when the real job is rethinking the experience for the new context. Covers assessing what breaks, phone-to-tablet restructuring, orientation and foldables, platform-to-platform idiom translation, and verifying custom controls under real touch input.
- ▌ Impeccable Audit · jzkk720Run systematic technical quality checks on a web interface and generate a scored report without fixing anything. Use when auditing implementation quality rather than critiquing design taste. Scores five dimensions 0-4 - accessibility, performance, responsive behavior, code quality, and theming - and documents findings for other procedures to address.
- ▌ Impeccable Shape · jzkk720Discover what should be made and how it should work, then return a confirmed design brief without writing code. Use before building a new surface or feature when the requirements are still open. Runs a bounded discovery interview on purpose, audience, material, and boundaries, resolves direction, then writes the smallest useful brief and stops.
- ▌ Loop Engineering · jzkk720Design, scaffold, score, schedule, and govern autonomous agent loops. Use when building loop infrastructure — the mechanism layer that runs, monitors, and recovers agent loops. Complements loopy (loop content/design) with loop execution infrastructure. Covers loop scaffolding, health scoring, cron scheduling, governance policies, and failure recovery patterns.
- ▌ Impeccable Harden · jzkk720Make an interface resilient to the unexpected: edge cases, hostile input, failure states, interrupted gestures, and real-world conditions. Use when an interface works only on the happy path, or before shipping anything users can break. Covers constraint handling, validation, accessibility and performance resilience, offline behavior, and the edge-case test matrix.
- ▌ Impeccable Android · jzkk720Android platform conventions for native apps built with Jetpack Compose, Android Views, React Native, Expo, or Flutter. Use when building or reviewing Android UI. Covers Material Design 3 as the governing rulebook, expressing brand through Material theming, the Android slop test for off-spec components, and the OS guarantees iOS hardware still owes.
- ▌ Impeccable Extract · jzkk720Identify reusable patterns, components, and design tokens in a codebase and consolidate them into the design system for systematic reuse. Use when repeated UI has drifted into near-duplicates, or when a design system exists but is being bypassed. Discovers the existing system first and refuses to invent a new one without being told where it belongs.
- ▌ Impeccable Operate · jzkk720Extended design depth for Operate-mode surfaces where the interface serves the product: app UIs, admin dashboards, settings panels, data tables, and authenticated tools. Use when the user is in a task rather than being persuaded. Covers the product slop test, earned familiarity as the bar, and the typography and consistency rules that Read surfaces share.
- ▌ Jev · jzkk720 bundleSet up Jev via OpenRouter, native TypeSafe or explicitly chosen simulation; build custom decision integrations and handle agent checkpoints, tool/model routing and context-retention decisions. Supply relevant context and batch independent questions. Focused collection skills cover bulk labels, evidence retrieval, evaluation, UI and simulations.
- ▌ Laya · jzkk720 bundleAdd fast, LOCAL, typed decisions to any project with Laya, an open-source non-generative decision model. Runs offline via `pip install laya`; classifies, routes, scores and answers yes/no questions about a piece of text, returning calibrated probabilities in ~20-35 ms on a laptop GPU, no LLM call, no data leaving the machine. Also use it when the user mentions Jev, a "System 1 model", "typed decisions" or a "decision model" for LOCAL/offline use, even if they never name Laya. Use when: routing text, triage tickets/emails, spam/phishing/toxicity/intent detection, guardrails in front of an agent or LLM, rubric scoring, replacing an LLM-based classifier to cut latency or cost. For HOSTED decisions use the `jev` skill instead; compose both via the cascade pattern in this skill.
- ▌
- ▌ Anti Slop · jzkk720Anti-AI-slop design and copy rules for any UI work (website, web app, or interface). Use whenever building, refining, or auditing user-facing UI. Holds every visual decision to a purpose test and the result to a liveliness bar so the output feels crafted by a designer, not generated by AI. Two modes: DURING (apply rules while building) or AFTER (audit finished work with a numbered findings list). Pairs with a DESIGN.md / brand direction for identity. Complements hallmark, taste-skill, baseline-ui, and improve-ui.
- ▌ Book To Skill · jzkk720Converts books and documents (PDF, EPUB, DOCX, HTML, Markdown, plain text, RTF, MOBI/AZW) into structured agent skills, extracting frameworks, mental models, principles, techniques, and anti-patterns. Use when the user wants to study a document through an agent, apply an author's frameworks while working, or build a reusable knowledge base from a file.
- ▌ Gstack Review · jzkk720Pre-landing PR review with structural-issue checklist and 8 specialist lenses. Use when asked to review a PR, do a code review, check a diff, or before landing/shipping code changes. Adapted from garrytan/gstack (MIT).
- ▌ Diagram Design · jzkk720Create branded architecture, IT current-state, flowchart, sequence, state machine, ER/data model, timeline, swimlane, quadrant, radar/spider, loop/ flywheel, nested, tree, org chart, layer stack, Venn, pyramid/funnel, bar, line, Gantt, scatter, high-level, process, medallion, data flow, DP integration, or DP security matrix diagrams as standalone HTML/SVG/PNG. Redraw .drawio/.drawio.png/.drawio.svg or Mermaid .mmd sources at a chosen size/detail; onboard brand tokens from a website; add semantic patterns, callouts, accessible motion, or sketchy/hand-drawn styling.
- ▌ Idea To Design · jzkk720You MUST use this before any creative work — creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements, and design through collaborative dialogue before any implementation. A pure-process methodology skill (no scripts, no shells, no browser servers). Adapted from the brainstorming skill methodology (obra/superpowers) with the unsafe visual-companion browser server removed.
- ▌ Webapp Testing · jzkk720Test local web applications with Playwright. Use when asked to verify frontend functionality, debug UI behavior, capture browser screenshots, or inspect browser logs. Pure-methodology skill: teaches the reconnaissance-then-action pattern and decision tree for static vs dynamic apps, with no bundled scripts. The agent writes native Playwright code or uses browser MCP tools instead.
- ▌ Open Code Review · jzkk720Performs AI-powered code review on Git changes using the `ocr` CLI from alibaba/open-code-review. Use when the user asks to review code, review a pull request, review staged/unstaged changes, review a commit, or compare branches for code quality issues. Produces line-level review comments and can automatically apply fixes when requested. With appropriate review rules, can detect various types of issues including bugs, security vulnerabilities, performance problems, and code quality concerns.
- ▌ Aris Idea Creator · jzkk720Generate and rank research ideas given a broad direction. Use when user says "brainstorm ideas", "generate research ideas", "what can we work on", or wants to explore a research area for publishable directions. Ported from wanshuiyin/Auto-claude-code-research-in-sleep (ARIS) — methodology only, Claude-Code-specific MCP/CLI/GPU deps stripped for VS Code Copilot.
- ▌ Aris Research Lit · jzkk720Search and analyze research papers, find related work, summarize key ideas. Use when user says "find papers", "related work", "literature review", "what does this paper say", or needs to understand academic papers. Ported from wanshuiyin/Auto-claude-code-research-in-sleep (ARIS) — methodology only, Claude-Code-specific MCP/CLI deps stripped for VS Code Copilot.
- ▌ Design Md Library · jzkk720Curated library of 74 real-world DESIGN.md files (Google Stitch spec) extracted from sites like Apple, Stripe, Linear, Vercel, Notion, Airbnb. USE WHEN the user wants a design system to ground UI generation, says "make it look like [brand]", "use the [brand] design system", wants to build a page matching a known site's visual language, asks for design tokens/colors/typography for a named brand, or references DESIGN.md / Stitch / VoltAgent awesome-design-md. Returns the path to the matching DESIGN.md so the agent can read and apply it. Not a generator itself — it's an index into the fork mirror at ~/dev/forks/JZKK720/awesome-design-md/design-md/.
- ▌ Impeccable Bolder · jzkk720Amplify a section within an existing design system to make it bolder and more confident. Use when a specific section reads flat or timid compared to its neighbors, and needs to be raised to the expressive level the rest of the page already reaches — without rebuilding anything outside the named target.
- ▌ Aris Novelty Check · jzkk720Verify research idea novelty against recent literature. Use when user says "novelty check", "check novelty", "has anyone done this", or wants to verify a research idea is novel before implementing. Ported from wanshuiyin/Auto-claude-code-research-in-sleep (ARIS) — methodology only, Claude-Code-specific MCP deps stripped for VS Code Copilot.
- ▌ Impeccable Animate · jzkk720Add purposeful motion to a user interface. Use when a design needs animation that explains state, relationship, and hierarchy — not decoration, but motion that carries meaning. Covers motion thesis, material-by-meaning, timing and easing, implementation strategy, and accessibility.
- ▌ Impeccable Clarify · jzkk720Improve UX copy, labels, error messages, and microcopy in a user interface. Use when interface text is unclear, ambiguous, jargon-heavy, or inconsistent — rewrite it so users understand what happened, what matters, and what to do next.
- ▌ Impeccable Delight · jzkk720Add personality and memorable moments to a product interface. Use when a design needs moments of joy, character, or emotional resonance — not generic whimsy, but product character revealed through useful interactions, humane responses, or unexpectedly considered details.
- ▌ Impeccable Distill · jzkk720Strip a design to its essence. Remove redundant elements, repeated information, decorative noise, and cosmetic complexity across information architecture, visual design, layout, interaction, content, and code. Use when a design feels cluttered, complex, or overloaded and needs simplification without losing necessary functionality.
- ▌ Impeccable Onboard · jzkk720Design first-run flows, empty states, and activation paths that get users to value as fast as possible. Use when building onboarding experiences, feature discovery, guided tours, interactive tutorials, or empty state design for a product interface.
- ▌ Open Code Review Delegate · jzkk720Delegation mode for open-code-review (OCR). Instead of OCR calling an LLM endpoint, this skill instructs the host agent to perform the code review itself, using OCR only for deterministic engineering: file selection and rule resolution. Use when the host agent should drive the review with its own LLM capabilities.