← all publishers

Kayforkind

@kayforkind source repo

417 published skills · page 1 of 5

  1. Reimagine It · kayforkind bundle
    Content-Derived Design — reads existing HTML and writes a stronger standalone page from headings, facts, names, dates, numbers, links, emails, and colors already in that file. CLI: npx reimagine-it --auto -i page.html -o out.html; also variations, lock, and audit. Use when the user says /reimagine-it, "reimagine it", "reinvent this", "redesign this page", "make an infographic", wants a content-aware redesign or a visual leap instead of a mood board. All 17 tokens: webpage, infographic, svg, 3js, simulation, artistic, cinematic, dashboard, photography, landing, editorial, gradient, showcase, motion, glass, lookbook, particles — plus leftover words as a brief. Also /reimagine-it audit for quality checks. The agent runs the CLI; it does not hand-author the page. Installed via npx skills add Kayforkind/reimagine-it or plugin marketplace. Not /better. Not a mood board. Does not ingest PDF, PPTX, or DOCX.
    0
    installs
  2. Reimagine It Lock · kayforkind
    Design DNA capture — extracts palette, type stack, motifs, motion, and 3D signatures from a shipped HTML file and saves them as a reusable reference pack. Use when the user says /reimagine-it lock, "lock this design", "save design DNA", "capture this look", or wants to reuse a shipped design on a different target. Locks capture palette, type, motif, and motion from a shipped HTML file. Part of the reimagine-it Content-Derived Design suite.
    0
    installs
  3. Reimagine It Audit · kayforkind
    Design Health — runs 19 deterministic quality checks on HTML output. Use when the user says /reimagine-it audit, "audit this page", "check design quality", or wants to verify craft-floor compliance before shipping. Catches blank plates, missing focus rings, non-compositor motion, off-palette accents, dead motion, fabricated content. No LLM, no API key. CI-ready with --json and exit codes. Part of the reimagine-it Content-Derived Design suite.
    0
    installs
  4. Reimagine It Extract · kayforkind
    Emit the content signals reimagine-it reads from an HTML file — title, anchors, proper nouns, dates, numbers, emails, links, source hex colors, and the derived palette — as JSON without generating a redesign. Use when the user says /reimagine-it extract, "what does the engine see in this page", "extract the palette", "pull the facts from this page", or wants content analysis for research, migration, or design-system work. CLI: npx reimagine-it extract -i page.html -o signals.json. Part of the reimagine-it Content-Derived Design suite.
    0
    installs
  5. Reimagine It Generate · kayforkind
    Generate one Content-Derived Design direction from an HTML source, or let Auto choose the strongest direction. Use when the user says /reimagine-it generate, "redesign this page", "make it a landing page", "turn this into an infographic", names a token (all 17: webpage, infographic, svg, 3js, simulation, artistic, cinematic, dashboard, photography, landing, editorial, gradient, showcase, motion, glass, lookbook, particles), or wants a content-aware redesign instead of a mood board. CLI: npx reimagine-it --auto -i page.html -o out.html. Part of the reimagine-it Content-Derived Design suite.
    0
    installs
  6. Reimagine It Variations · kayforkind
    Generate several distinct design directions from one HTML source plus a contrast sheet that shows every direction side by side. Use when the user says /reimagine-it variations, "show me options", "give me three takes", "which direction works best", or wants to compare design directions before choosing one. CLI: npx reimagine-it variations -i page.html -n 4 -o review/. Part of the reimagine-it Content-Derived Design suite.
    0
    installs
  7. Reimagine It Infographic · kayforkind
    Content-derived infographic poster — reads the source file and builds a statistical poster from the facts already in it. Not a dashboard, not a CV, not an AntV template. Uses common-scale encodings (Cleveland–McGill floor), ISOTYPE unit counts (Neurath), custom data glyphs (Lupi/Fragapane), and a lossless data table. Palette, glyphs, and composition are derived from the source's own content. No pies, donuts, gauges, 3D charts, or fabricated KPIs. Part of the reimagine-it Content-Derived Design suite.
    0
    installs
  8. Client View · kayforkind bundle
    Embeds Zoom Meeting SDK Web Client View through the ZoomMtg singleton: system check, WASM preload, init/join with passWord (capital W), and inMeetingServiceListener. Use for full-page Zoom meeting embeds or debugging blank-screen, black-overlay, and join-payload failures. Not for Zoom Component View custom tile layouts or Zoom Video SDK canvases.
    1
    install
  9. Examprep AI · kayforkind
    Converts syllabi, past papers, or notes into an Easy-to-Medium-to-Hard High Score Roadmap plus flashcards, predicted papers, and readiness checks. Use when students ask for exam prep, revision plans, question prediction, or mock papers. Not for sitting a live exam or unrelated homework tutoring; never reverse the Easy then Medium then Hard study order.
    1
    install
  10. Expo Module · kayforkind bundle
    Builds Expo native modules and views with the Expo Modules API (Swift, Kotlin, TypeScript), including create-expo-module, expo-module.config.json, config plugins, and lifecycle hooks. Use when adding native functionality, native views, or platform support to an Expo module. Not for Expo Router screens, JS-only Expo app features, or React Native Turbo Modules outside the Expo Modules API.
    1
    install
  11. Fp Refactor · kayforkind
    Migrates existing imperative TypeScript—hidden throws, nullable returns, and Node err-first callbacks—onto fp-ts Option, Either, and TaskEither so callers must handle absence and failure. Use when throws, nulls, or callbacks hide failure from the compiler and a staged conversion is justified. Not for greenfield Either snippets or profiler-hot inner loops. Never convert stable readable code solely for style, or introduce fp-ts where the team cannot maintain it.
    1
    install
  12. Gen Catalog · kayforkind
    Indexes SKILL.md trees by reading name and description frontmatter, inferring categories, and emitting README marker tables, JSON, or standalone HTML. Use when generating or refreshing a skills catalog, README skills table, or skills-list embed. Not a /create-skill authoring or quality-refresh pass. Never invent missing YAML fields, nest one skill under another, or add Co-Authored-By on the optional catalog commit.
    1
    install
  13. Google Vids · kayforkind
    Creates structured Google Vids (vids.new) work videos via Help me create: prompt plus attached Docs/Sheets/Slides become outline, scenes, stock media, script, AI voiceover, 8s Veo clips, and avatars. Use when the user wants an explainer, launch, or training video from a prompt or document. Not for cinematic Veo/Flow footage, Slides-only decks, or live-action NLE editing.
    0
    installs
  14. Grokimagine · kayforkind bundle
    Grok Imagine via SuperGrok Heavy in real Chrome (Playwright CDP). Prefer Video I2V with parallel fire (submit→close tab→next; harvest later); Image locks; Agent films; UUID export. Trigger on /grokimagine. Never Cursor IDE browser. Ask before xAI API credits. Existing character films: continue the Agent conversation — never New Generation / home Video (face drift).
    0
    installs
  15. Manim Video · kayforkind bundle
    Produces Manim Community Edition explainer videos: Scene classes, MathTex, ffmpeg stitch, optional TTS — 3Blue1Brown-style math, algorithms, and architecture diagrams. Use when the user wants programmatic math/technical animation with Manim. Not for Remotion/React video (remotion-video), character cinema (Blender/AE), or interactive apps.
    1
    install
  16. Mesh Memory · kayforkind
    Runs self-hosted semantic agent memory over PostgreSQL/pgvector via MCP tools (mesh_add, mesh_search, mesh_focus) with local e5 embeddings. Use when persisting worklogs, decisions, and notes for cross-session recall by meaning. Not for Chroma RAG stores (chroma), local markdown search with qmd, or Obsidian vault note CRUD (obsidian).
    1
    install
  17. Modern HTML · kayforkind
    Authors and reviews 2026 Baseline HTML: document head, semantic landmarks, native dialog/popover/invoker commands, web components, forms, WCAG 2.2, JSON-LD, loading/CLS, and HTML security (CSP/SRI). Use when writing or auditing markup rather than CSS or framework state. Not for CSS layout/animation (modern-css), React/Vue architecture, or bundler config.
    1
    install
  18. PPTX Author · kayforkind
    Writes .pptx files headless with python-pptx (pitch decks, IC memos, earnings notes), optionally binding every figure to an Excel model via openpyxl. Use when the deliverable is a .pptx on disk, especially model-backed finance decks paired with excel-author. Not for live PowerPoint/Office MCP sessions or animation-heavy all-hands decks (powerpoint). Never email or upload the file from this skill — write ./out only.
    1
    install
  19. Statsmodels · kayforkind bundle
    Fits statsmodels OLS/WLS/GLS, GLM families, logit/probit/MNLogit, and ARIMA/SARIMAX/VAR with inference, sandwich SEs, and residual tests. Use for publication tables, forecast intervals, or assumption checks (Breusch-Pagan, ADF, Cook's distance). Not for scikit-learn pipelines or NumPy-only numerics; never a symbolic-algebra chair (sympy).
    1
    install
  20. Svg Creator · kayforkind
    Creates, edits, sanitizes, and packages spec-correct standalone SVG for icons, logos, diagrams, and SMIL animation. Use when generating or fixing .svg markup or converting a visual concept to SVG. Not for corpus quality audits (svg-quality-audit), React/Tailwind UI (design-taste-frontend), or raster PNG illustration.
    1
    install
  21. Test Fixing · kayforkind
    Groups failing suite errors by type and root cause, then repairs infrastructure, API drift, and assertion bugs until the project's runner is green. Use when tests fail, CI is red, or a refactor broke the suite. Do not use to author a new pytest or TestNG tree, or to skip or weaken assertions for a pass.
    1
    install
  22. Trigger Dev · kayforkind
    Defines Trigger.dev v3 TypeScript tasks: durable background jobs, cron, webhooks, batches, retries, and trigger.config.ts at the package root. Use when the user mentions Trigger.dev, AI/background jobs past serverless timeouts, or Trigger cron/webhooks. Not for BullMQ, Inngest, Temporal sagas, or Wrangler Cron Triggers. Never hardcode TRIGGER_SECRET_KEY or let SDK and CLI versions drift.
    1
    install
  23. YAML Config · kayforkind
    Reads and writes PyYAML configs with yaml.safe_load, utf-8 pathlib I/O, YAMLError line numbers, and factory-defaults deep-merge for vehicle or app parameters. Use when parsing config.yaml, dumping with sort_keys=False, or merging user overrides. Never call yaml.load without a Loader; not for GitHub Actions workflow YAML or multi-gigabyte JSONL/Parquet dumps.
    1
    install
  24. Agents Sdk · kayforkind
    Build, run, deploy, and evaluate OpenAI Agents SDK apps. Use when the user asks to create or adapt an Agents SDK app, build from a prompt or Codex thread, prepare a runnable agent prototype, add a focused eval harness, or deploy locally through the Agents SDK Deployment Manager. Not for LangChain/CrewAI or generic Chat Completions wrappers without the OpenAI Agents SDK.
    1
    install
  25. Aiq Deploy · kayforkind bundle
    Deploys, validates, troubleshoots, and stops NVIDIA AI-Q Blueprint 2.1.x (Docker Compose, Kubernetes/Helm, local process, CLI, or web UI). Use when installing or cloning AI-Q, starting the backend, checking /health, fixing port or credential failures, or handing a verified AIQ_SERVER_URL to aiq-research. Not for deep research report generation (that is aiq-research).
    1
    install
  26. Bug Hunter · kayforkind bundle
    Hunts bugs with reproduce-then-evidence debugging: logs, bisect, and a regression test before claiming a fix. Use when the user reports a crash, error, flake, or asks to debug or fix a bug. Not for greenfield feature design. Do not use as a substitute for the project's test runner or typecheck.
    1
    install
  27. Bulletmind · kayforkind bundle
    Rewrites any input as a hyphen-only hierarchical bullet tree (lite, full, or ultra) with 2-space nesting and no prose bridges. Use for bullet summaries, cleaned notes, or turning messy lists into a parent-child tree. Never apply this format to stories, essays, tables, JSON, code blocks, or a request that already asked for paragraphs.
    1
    install
  28. Clerk Auth · kayforkind
    Implements Clerk on Next.js App Router: ClerkProvider, clerkMiddleware/auth.protect(), organizations, Svix user-sync webhooks, and server auth()/client hooks. Use when the user mentions Clerk, ClerkProvider, sign-in/sign-up, organizations, SSO, or Clerk webhooks on Next.js 14/15. Not for Auth.js, NextAuth, Firebase Auth, or Clerk on non-Next stacks. Never trust middleware alone (CVE-2025-29927); re-check auth() in the handler.
    1
    install
  29. Fast Check · kayforkind
    Adds JavaScript and TypeScript property-based tests with fast-check (randomized inputs, shrinking, roundtrips, custom arbitraries, fc.assert). Use when verifying invariants across generated inputs rather than single examples. Not for Vitest/Jest checks of fixed API fixtures, component rendering, or process-crash fuzzing.
    1
    install
  30. Game Godot · kayforkind bundle
    Tests, exports, and deploys Godot 4 games with GdUnit4 (in-engine) and PlayGodot (external Python E2E). Use when writing GdUnit4 tests, PlayGodot automation, or Godot --export-release CI. Not for GDScript style (godot-gdscript-mastery) or Control/Theme UI (godot-ui). Never vercel deploy --csp — that flag does not exist.
    1
    install
  31. Localvideo · kayforkind bundle
    Local LongCat-Video on Windows (high-VRAM NVIDIA) with identity-first flow (hero still → I2V → lengthen → upscale). Use for /localvideo; never text-only T2V for celebrity/character likeness finals.
    1
    install
  32. Matplotlib · kayforkind
    Builds Matplotlib figures with the OO Axes API: subplots, mosaic, GridSpec, rcParams, and savefig to PNG/PDF/SVG. Use for pyplot line, scatter, bar, hist, heatmap, contour, or 3D work in Python. Never route Seaborn estimators or Plotly Dash apps through this chair.
    1
    install
  33. Modern CSS · kayforkind
    Writes Baseline 2026 native CSS: Flexbox/Grid/subgrid, logical properties, container queries, @layer/@scope, :has, oklch, nesting, anchor positioning, and fluid type. Use when writing or reviewing stylesheets, layout, dark mode, or cascade without a preprocessor. Not for Tailwind v4 @theme setup (setup-tailwind-typescript), glassmorphism recipes, or Sass/CSS-in-JS. Never default to physical margin-left/width when logical properties apply.
    1
    install
  34. Page Agent · kayforkind
    Embeds alibaba/page-agent as an in-page GUI agent (DOM-as-text, OpenAI-compatible LLM) so end users drive a site with natural language. Use when adding an NL copilot inside a SaaS/admin app or evaluating page-agent against Ollama, Qwen, OpenAI, or OpenRouter. Not for an agent driving a browser from outside (Playwright/Browserbase) or screenshot/multimodal grounding. Never ship the demo CDN LLM proxy to production.
    1
    install
  35. Rclone CLI · kayforkind bundle
    Runs rclone for cloud file ops across 70+ remotes: copy/sync/move, mount, bisync, encrypted backups, server-side copy, and S3-compatible config. Use when the user mentions rclone or terminal cloud transfer, mount, or remote management. Not for vendor web consoles or treating rclone.conf as shareable; never log rclone config show with redaction disabled.
    2
    installs
  36. Statusline · kayforkind
    Configures the Cursor CLI status line (statusLine.command in cli-config.json): stdin JSON for model, context %, git/worktree, ANSI stdout above the prompt. Use when the user mentions statusline, statusLine, CLI status bar, prompt footer, or a session context bar. Not for Starship, oh-my-posh, VS Code status bar, or tmux. Never block longer than timeoutMs; the line is local and does not spend API tokens.
    1
    install
  37. Transition · kayforkind bundle
    Remotion, CSS, and ffmpeg transition recipes (whip pans, zoom ramps, glitch cuts, film burns, parallax) for quote films and Shorts. Use when building QuoteLong/QuoteShort beat-synced cuts or caption motion language. Not for Remotion project setup and render (remotion-video) and not the /effects content-type matrix (inspire-quote-effects).
    1
    install
  38. UI Toolkit · kayforkind bundle
    Integrates Zoom Video SDK UI Toolkit as drop-in gallery, controls, chat, and screen-share UI for custom sessions. Use when embedding branded video calling inside a web app without building conference chrome. Not for joining real Zoom meetings (component-view / Meeting SDK) and not for raw Video SDK custom rendering.
    1
    install
  39. Qmd · kayforkind
    Indexes local notes, docs, and transcripts with qmd hybrid retrieval (BM25 + vectors + LLM rerank) via @tobilu/qmd. Use when searching a personal knowledge base semantically without cloud APIs. Not for Obsidian vault editing (obsidian), Mesh Memory MCP worklogs (mesh-memory), or arXiv paper search (arxiv).
    1
    install
  40. Clip · kayforkind bundle
    Runs OpenAI CLIP for zero-shot image classification, image-text cosine matching, semantic search, and cross-modal retrieval in a shared embedding space. Use when labeling images without training, ranking images by a text query, or scoring vision-language similarity. Not for captioning (BLIP-2), conversational VLM chat (LLaVA), or pixel-level segmentation (SAM); never skip L2-normalizing embeddings before cosine scores.
    1
    install
  41. Crit · kayforkind
    Drives the crit CLI blocking-review loop: GitHub-style inline comments on diffs, plans, live URLs, or local HTML, then applies unresolved review.json comments; also covers UI/UX accessibility and responsive design review. Use when the user asks to review, critique, revise UI, or run crit. Not for /better quality passes, GitHub PR reviews that never launch crit (github-code-review), or WCAG-only audits.
    1
    install
  42. Goal · kayforkind bundle
    Outcome contract then autonomous loop. Use when the user runs /goal, /goal-loop, “capture goal”, “GOAL.md”, “until done”, or the work spans many files/steps and success must be falsifiable. Writes GOAL.md before wandering. /goal-loop is a mode of this skill (OBSERVE→PLAN→ACT→VERIFY→RECORD), not a third always-on skill. Not /loop (session interval ticks). Not /better (quality pass on an existing deliverable).
    1
    install
  43. Loop · kayforkind
    Arms a Cursor-session interval loop: parse /loop [interval] <prompt>, emit AGENT_LOOP_TICK sentinels from a monitored PowerShell/bash shell, run the prompt on each tick, and stop by PID. Use when the user says /loop, every 5m, or keep checking. Not for /goal-loop GOAL.md autonomy, OS crontab, or one-shot tasks.
    1
    install
  44. Peft · kayforkind bundle
    Fine-tunes LLMs with Hugging Face PEFT (LoRA, QLoRA, IA3, multi-adapter serve) training under 1% of weights. Use when adapting 7B–70B models on limited GPU memory or serving multiple adapters from one base. Not for full-weight fine-tuning of small models and not for hosting Gradio demos (huggingface-spaces).
    1
    install
  45. Plan · kayforkind
    Writes a bite-sized, copy-pasteable implementation plan to .hermes/plans/ with no code execution. Use when the user says plan, /plan, break down, or wants a roadmap before coding. Not for GOAL.md autonomous loops (goal) or immediate implementation. Do not run mutating commands, commits, or installs while planning.
    1
    install
  46. Snyk · kayforkind
    Runs Snyk CLI audits with snyk test (SCA), snyk code test (SAST), snyk container test, snyk iac test, and optional snyk monitor or GitHub Actions gates. Use for lockfile CVEs, image OS packages, or Terraform/Kubernetes misconfig. Never write Semgrep YAML here; not a pre-commit formatter or generic Node project-setup chair.
    1
    install
  47. Svgo · kayforkind
    Runs SVGO v3 with ESM config to strip Inkscape/Illustrator namespaces, clean paths, and batch-compress SVG folders while keeping viewBox and interactive IDs. Trigger on optimize-svg, icon minification, or a CI dry-run gate. Do not use for PNG/JPEG/WebP rasters or CSS layout work.
    1
    install
  48. Vllm · kayforkind bundle
    Serves LLMs with vLLM PagedAttention and continuous batching: OpenAI-compatible /v1/chat/completions, GPTQ/AWQ/FP8 quantization, tensor parallelism, and offline LLM.generate. Use when deploying high-throughput GPU inference endpoints or fitting 30B-70B models. Not for TRL training, llama.cpp CPU/edge, or using vLLM only as an inspect-ai/lighteval backend.
    1
    install
  49. Email · kayforkind
    Sends transactional and broadcast email from Next.js/Vercel apps with Resend and React Email templates (domains, batch, webhooks, idempotency). Use when installing resend or react-email, writing React Email templates, or building app/api/send routes. Not for IMAP mailbox CLI (himalaya). Never expose RESEND_API_KEY to the client.
    1
    install
  50. Pymol · kayforkind bundle
    Visualize, analyze, and render protein and molecular structures using PyMOL. Use when the user wants to create images of protein structures, perform structural alignments or superposition, measure distances or contacts, highlight binding sites or active site residues, color by B-factor/pLDDT, or analyze protein-ligand interactions. Do not use for docking, molecular dynamics, or sequence-only analysis.
    1
    install
  51. Qutip · kayforkind bundle
    Simulates quantum states and open systems in QuTiP 5.x (sesolve, mesolve, mcsolve, entanglement, Bloch/Wigner, Floquet, HEOM). Use when evolving Qobj states, measuring fidelity or entropy, or visualizing Bloch/Wigner plots. Not for classical ODEs, large-scale circuit transpilation (qiskit/cirq), or loading untrusted .qu pickles.
    1
    install
  52. Rdkit · kayforkind bundle
    Drives RDKit Chem.Mol for SMILES/SDF I/O with sanitization control, descriptors, Morgan fingerprints, SMARTS substructure, reaction SMARTS, and seeded ETKDG 3D embed. Use for custom sanitization or algorithms datamol hides. Do not use for mordred million-row descriptor mills, quantum energetics, or unsandboxed attacker-controlled MOL/SDF.
    1
    install
  53. Slack · kayforkind bundle
    Routes Slack workspace tasks: search and read channels, threads, DMs, and canvases, then hand off composition. Trigger on Slack recap, triage, search, or post intent. Never finalize outbound mrkdwn here (slack-outgoing-message) and not for channel-create, message-edit, delete, or admin APIs.
    1
    install
  54. Sympy · kayforkind bundle
    Computes exact SymPy results: assumed symbols, solve/solveset/dsolve, calculus, symbolic matrices, lambdify, and LaTeX or C/Fortran codegen. Use when the user needs closed-form algebra instead of floats. Not for NumPy/SciPy heavy numerics or Monte Carlo integration.
    1
    install
  55. Jq · kayforkind
    Writes jq filters to query, reshape, aggregate, and pipeline JSON from APIs, CLIs, and logs. Use when parsing JSON, building jq one-liners, or explaining a jq expression. Not for in-memory DataFrames (pandas/polars), JSON Schema validation, or JavaScript Array.map as a jq substitute.
    1
    install
  56. Applicationinsights Web TS · kayforkind
    Instrument browser/web apps with @microsoft/applicationinsights-web for Real User Monitoring (RUM), page views, clicks, AJAX/fetch dependencies, exceptions, custom events, and browser-side GenAI agent traces. Use when adding App Insights to React, Angular, Vite, Next.js, or React Native frontends. Not for Node.js server OpenTelemetry (azure-monitor-opentelemetry-ts).
    1
    install
  57. Consistent Characters 2026 · kayforkind
    Locks a still-image identity with a six-tier ladder from a single canonical hero frame: verbatim tokens, Midjourney --cref/--cw and v7 --oref/--ow, InstantID/PuLID, Flux Kontext/USO star edits, then a character LoRA, gated by a contact sheet. Use for comics, mascot pose packs, turnarounds, or NPC expression grids. Never treat this as a video/ComfyUI animation graph (consistent-character-builder); not a style-only --sref chair or a rigged-3D identity source.
    1
    install
  58. Drizzle Migration Conflict · kayforkind bundle
    Classifies and repairs Drizzle Kit clashes in _journal.json, snapshots, and generated SQL after pull, merge, or rebase, including non-commutative drizzle-kit check failures. Use when migration artifacts conflict or a team needs regenerate-from-merged-schema plus CI policy. Do not use for schema modeling, drizzle-kit push as the production fix, or migrate against an unconfirmed database.
    1
    install
  59. Ffmpeg Keyframe Extraction · kayforkind
    Extract keyframes (I-frames) from video as still images with FFmpeg using select filter or -skip_frame nokey, with PTS-stamped filenames and ffprobe keyframe counting. Use when pulling thumbnails, scene-representative frames, or ML sampling frames out of MP4/MKV/AVI/WebM. Not for cutting or joining clips, scaling/watermarks/effects, codec or container changes, audio work, or reading stream metadata.
    1
    install
  60. Game Performance Profiling · kayforkind
    Use when a game drops frames, stutters, or must hit a frame budget — engine-agnostic profiling and optimization: CPU-vs-GPU bottleneck diagnosis, millisecond frame budgets, draw-call batching/instancing, overdraw and fill rate, LOD and culling, GC-free code and object pooling, memory layout (AoS vs SoA, cache misses), job/worker systems, and capture tools (Unity Profiler, Unreal Insights/stat, RenderDoc, PIX, Tracy, Superluminal). Triggers on profiling, frame budget, bottleneck, CPU bound, GPU bound, draw calls, batching, instancing, overdraw, fill rate, LOD, culling, object pooling, garbage collection, cache miss, SoA, job system, RenderDoc, Tracy. Not for Godot-editor profiling workflow (use game-godot-performance-optimization), Godot debugger/breakpoint diagnostics (use game-godot-debugging-profiling), authoring shaders/VFX (use technical-art-vfx), or server tick and bandwidth budgets (use game-multiplayer-netcode).
    1
    install
  61. Locational Marginal Prices · kayforkind
    Extracts locational marginal prices (LMPs) as DC-OPF nodal-balance duals (CVXPY + CLARABEL/OSQP), plus reserve MCP. Use for nodal electricity prices, congestion, or transmission counterfactuals. Not for AC-OPF (voltage/Q), unsolved duals, inventory EOQ (inventory-demand-planning), or NVIDIA cuOpt routing (cuopt-user-rules).
    1
    install
  62. Performance And Web Vitals · kayforkind
    Remediates Core Web Vitals in page assets and markup: LCP preload and fetchpriority, CLS width/height, INP main-thread splits, WebP/AVIF, and font-display. Use for LCP above 2.5s, CLS above 0.1, or INP above 200ms in the site itself. Never substitute web-vitals-analyzer (report RCA) or frontend-lighthouse (CI budget gates).
    1
    install
  63. Browser Extension Builder · kayforkind
    Scaffolds Manifest V3 Chrome, Firefox, and cross-browser extensions: content scripts, service workers, popups, chrome.storage, monetization, and Web Store/AMO publishing. Trigger on browser extension, chrome extension, firefox addon, MV3, or extension popup work. Do not use for Electron apps, PWAs, or ordinary websites unless the deliverable is an unpacked or store extension.
    1
    install
  64. Browser Game Architecture · kayforkind bundle
    Owns the HTML5/itch.io game shell: fixed-timestep loop, scene stack, input, audio unlock, asset preload, versioned saves, letterbox scaling, and Vite/PWA build. Use when starting or restructuring a Canvas, WebGL, or DOM-hybrid web game. Not for Three.js renderer APIs (threejs-game-development), Godot/Unity, or generic SPA dashboards.
    1
    install
  65. Chrome Browser Automation · kayforkind bundle
    Agent-driven Chrome automation for research, web apps, downloads, screenshots, and game/canvas testing via owned CDP Chrome + Playwright. Use for browser control, CDP 9222, multi-tab workflows, form fill, download, screenshot verify. Compose with a dedicated Chrome user-data dir over CDP 9222. Never invent profiles. Never cursor-ide-browser for Grok/Flow auth. Not for CAPTCHA bypass or anti-bot evasion.
    1
    install
  66. Designing Workflow Skills · kayforkind
    Architects workflow SKILL.md files with numbered phases, routing/safety-gate/task-driven patterns, tool assignment, and progressive disclosure. Use when building or reviewing a multi-step skill structure rather than its domain content. Not for Cursor skill scaffolding (create-skill) or hermes-agent in-repo authoring (hermes-agent-skill-authoring).
    1
    install
  67. Godot Debugging Profiling · kayforkind bundle
    Profiles and debugs Godot 4.7 with the Debugger, breakpoint, push_error, Performance.get_monitor, orphan nodes, remote_debug, and release-export timing. Use when the user wants Godot-specific leak hunts, draw-call spikes, or editor profiler frame time. Trigger: print_debug, assert, OBJECT_ORPHAN_NODE_COUNT, Time.get_ticks_usec. Not for engine-agnostic git bisect or regression hunting (bug-hunter). Never profile in a debug build or with V-Sync on, and never leave unlabeled print() in release.
    1
    install
  68. Inventory Demand Planning · kayforkind bundle
    Chooses retailer forecast families (moving average, Holt-Winters, causal, LightGBM), safety stock under lead-time risk, reorder/EOQ/min-max, and promo lift with cannibalization and post-promo dip. Trigger on demand forecast, safety stock, replenishment, or promo lift across stores and SKUs. Not for factory MRP, 3PL slotting, or warehouse WMS build-out.
    2
    installs
  69. Mvp Scoping And Risk Test · kayforkind
    Defines the smallest product slice and cheapest experiment to test a startup's core value hypothesis before heavy engineering. Use when a founder is scoping an MVP, validating demand, designing a smoke/concierge/Wizard-of-Oz test, or fighting over-engineering. Not for post-PMF growth, live A/B statistical readout, or writing the production implementation; never treat opinions as a substitute for behavioral pass/fail bars.
    1
    install
  70. Postman Newman Automation · kayforkind
    Emits Newman CLI invocations, PowerShell or bash runners, and Jenkins/GitHub Actions stages for exported Postman collections, including CSV iteration data and JUnit/htmlextra reporters. Trigger on newman, postman cli, collection CI, or data-driven API suites. Do not use for Postman GUI collection authoring, the Postman CLI binary, Bruno/Insomnia clients, or Pact contract tests.
    1
    install
  71. Remotion Video Generation · kayforkind
    Structures Remotion 4.x React compositions including fps, durationInFrames, Zod props, frame-accurate timelines, CLI and Lambda render, fonts, and remote-asset CORS. Use when building, configuring, or debugging Remotion compositions or render pipelines. Not for quote-film transition recipes (transition), ComfyUI or LongCat generation, or ffmpeg-only transcodes.
    1
    install
  72. Setup Tailwind Typescript · kayforkind
    Scaffolds Tailwind CSS v4 in a TypeScript app (Next.js, Vite, or React) via @tailwindcss/postcss, CSS-first @theme tokens, cn() (clsx + tailwind-merge), and class-based dark mode. Use when adding Tailwind v4, a css-first theme, or a dark-mode toggle to a TS project. Not for Tailwind v3 tailwind.config.js, native CSS without Tailwind (modern-css), or glassmorphism recipes. Never run npx tailwindcss init — v4 dropped it and the content array.
    1
    install
  73. Social Metadata Hardening · kayforkind
    Hardens Open Graph and Twitter Card tags so public URLs unfurl as rich cards on Facebook, LinkedIn, X, WhatsApp, Telegram, Slack, and Discord. Use when shared links show missing, stale, cropped, or broken previews, or when auditing og and twitter metadata before launch. Not for paid social ad creatives or sitemap-only SEO; never rely on client-side JS to inject tags crawlers must see.
    1
    install
  74. State Space Linearization · kayforkind
    Linearizes nonlinear plants to local (A,B) via Jacobian or autodiff, then ZOH/Tustin discretize for LTI design. Use when a nonlinear model must feed LQR, MPC, Kalman, or gain scheduling. Not for finite-horizon LQR itself (finite-horizon-lqr), hard discontinuities, or naive Euler on stiff dynamics.
    1
    install
  75. Storyboard Video Pipeline · kayforkind
    Assembles a final cut from a storyboard manifest plus per-shot clips and audio: timed concat, audio mux, SRT burn, and verification frames via ffmpeg. Use when the request names manifest, clips, and audio together and asks to stitch, mux, or burn. Not for bare concat or mux without a manifest, generative T2V prompting, or NLE timeline editing; never treat cmd /c as safe for filter_complex on Windows.
    1
    install
  76. Timed Lyric Transcription · kayforkind
    Produces word- and segment-timed transcripts with WhisperX or faster-whisper, optional Demucs vocal stems, then SRT, VTT, ASS karaoke, or JSON for lip-sync. Trigger on timestamped lyrics, karaoke from MP3, force-align lyrics, viseme tracks, or Whisper JSON to SRT. Do not use for livestream captioning or DRM-encrypted audio.
    1
    install
  77. Agent Prompt Engineering · kayforkind
    Load when writing prompts for sub-agents, generation CLIs (jules, agy, claude-code, opencode, codex), or any model emitting structured artifacts (SVG, JSON specs/ops, code files). Turns vague instructions into machine-checkable contracts with mandatory verification gates. Not for casual chat with no artifact deliverable, and not for authoring a new skill file (create-skill).
    1
    install
  78. Angular State Management · kayforkind bundle
    Guides Angular 16+ state: Signals and signal services, NgRx SignalStore and Store (actions, reducers, effects), ComponentStore, and toSignal/toObservable bridges. Use when choosing Signals vs NgRx vs Akita, migrating BehaviorSubject stores, or implementing optimistic server state. Not for React, Vue, or AngularJS state libraries. Do not use when the only task is wiring HttpClient.
    1
    install
  79. Comfyui Prompt Interview · kayforkind bundle
    Interviews a vague image idea across 4-7 exchanges, then synthesizes a model-tuned positive prompt, negative prompt, settings table, and pipeline recommendation. Use when the user has a fragment vision and needs a ComfyUI prompt spec before generation. Not for ComfyUI node-graph wiring, repairing an existing underperforming prompt, LoRA training, or claiming an image was rendered.
    1
    install
  80. Comfyui Workflow Builder · kayforkind bundle
    Use when generating, building, creating, designing, or repairing ComfyUI API-format workflow JSON from a natural-language description (txt2img, img2img, inpainting, ControlNet, LoRA stacking, upscaling, face detailing, FLUX, Wan I2V). Load when the user asks for a node graph, pipeline JSON, or runnable ComfyUI prompt that must use correct class_types, connection wires [nodeId, outputIndex], inventory-backed model filenames, and model-native resolutions. Do not use for ComfyUI install/server setup, custom node development, Python scripting, model training, GPU buying advice, or pure architectural explanations of diffusion.
    1
    install
  81. Contract Playbook Review · kayforkind
    Compares each YAML playbook rule to the matching contract provision and emits one ordered finding with found, character-exact excerpt, and status derived from action_* (ok/risk/reject). Apply when an NDA, MSA, DPA, lease, or vendor DD questionnaire must be scored against machine-checkable fields such as max_years or acceptable_jurisdictions. Not a contract-drafting generator and not advice beyond encoded rules; skip reviews that have no playbook.
    1
    install
  82. Customs Trade Compliance · kayforkind bundle
    Walks HS/HTS/TARIC classification (GRI 1-6, CROSS/BTI), FTA origin and RVC, hierarchical customs valuation, Incoterms, and restricted-party screening (SDN, Entity List, Denied Persons). Load when opening a new import/export lane, checking duty/FTA/FTZ/drawback, or adjudicating a screening hit. Never classify from a SKU name; not for domestic sales-tax engines or warehouse WMS build-out.
    1
    install
  83. Excitation Signal Design · kayforkind
    Design the excitation experiment for system identification — pick the input signal (step, multi-level step, PRBS, multisine, chirp), size the step magnitude, test duration, and sample rate, and collect a clean input/output record from a plant you can safely perturb. Triggers on 'step test', 'excitation signal', 'identification experiment', 'how long should the test run', 'what sample rate for identification'. Not for fitting parameters to an already-recorded response (use first-order-model-fitting) or turning an identified model into controller gains (use imc-tuning-rules).
    1
    install
  84. Ffmpeg Format Conversion · kayforkind
    Convert media between containers and codecs with FFmpeg - remux with -c copy, transcode to H.264/HEVC/AV1/Opus, CRF vs two-pass bitrate, hardware encoders (NVENC/QSV/VideoToolbox). Use when changing file format (MKV to MP4, WAV to Opus), shrinking files, or fixing playback compatibility. Not for trimming/joining clips (use ffmpeg-video-editing), visual effects (ffmpeg-video-filters), loudness/mixing (ffmpeg-audio-processing), or inspecting streams (ffmpeg-media-info).
    1
    install
  85. Game Creativity Director · kayforkind
    Directs the pre-production vision layer of a game: player personas, market positioning, feeling-led pillars, volume ideation, cheapest-prototype choice, toy tests, level pitches, and a style bible, then hands off. Use when the work is still exploring identity, audience, and fun before a build. Not for loop tuning, dialogue topology, shader authoring, or engine code. Never treat this chair as a Godot or Unity implementation skill or a substitute for gameplay-and-design.
    1
    install
  86. Game Godot Genre Romance · kayforkind bundle
    Architects Godot 4 romance and dating-sim loops: multi-axis affection (attraction/trust/comfort), signal-driven TimeManager, date events, and route branching. Use when building otome, Persona-style social links, or relationship-centric Godot games. Not for other Godot genres, binary love/hate meters, action/puzzle games without romance systems, or Dialogic dialogue-box rendering.
    1
    install
  87. Game Godot Genre Shooter · kayforkind bundle
    Godot 4 FPS/TPS gunplay blueprint covering WeaponData resources, recoil and bloom, hitscan vs projectile, aim assist, client prediction, and hit registration. Use when building competitive shooters, battle royales, or tactical FPS in Godot 4. Not for platformers (game-godot-genre-platformer), engine-agnostic NPC AI (game-ai-behavior), GdUnit export CI (game-godot), or Unity/Unreal shooters.
    1
    install
  88. Game Multiplayer Netcode · kayforkind
    Use when designing or debugging real-time netcode in any engine — choosing a sync model (authoritative server vs deterministic lockstep vs rollback/GGPO), client-side prediction with server reconciliation, snapshot interpolation, lag-compensated hit registration, and bandwidth work (delta compression, quantization, interest management). Triggers on 'rubber-banding', 'desync', 'hit reg', 'client prediction', 'reconciliation', 'rollback', 'GGPO', 'tick rate', 'lag compensation', 'jitter buffer'. Not for Godot connection plumbing — peers, @rpc semantics, node authority (use game-multiplayer-basics), MultiplayerSynchronizer setup and tuning (use game-multiplayer-sync), headless/dedicated server builds (use dedicated-server), or lobbies and matchmaking services (use game-steamworks-sdk).
    1
    install
  89. Godot Scene Organization · kayforkind
    Organizes Godot 4.4+ scene trees around one-concept scenes, composition versus inherited variants, split-versus-keep heuristics, and signals-up / calls-down / EventBus-sideways communication. Use when choosing hierarchy, extracting a sub-scene, or wiring HUD and level instances. Not a component-class implementation chair (godot-component-system) and not for throwaway prototype scenes.
    1
    install
  90. Odoo Inventory Optimizer · kayforkind
    Configures Odoo 16+ Inventory: per-category FIFO or AVCO automated valuation, min/max replenishment, putaway locations, 2- and 3-step warehouse routes, and lot/serial tracking. Use when negative stock, broken cost layers, or Pick-Pack-Ship is the work. Do not use for QWeb report XML or purchase-pricelist setup.
    1
    install
  91. Photopea Embedded Editor · kayforkind
    Embeds Photopea in a host web app via photopea.js (createEmbed, runScript, saveToOE) for PSD-like layers, text, filters, file I/O, and Photoshop-compatible scripts. Use when integrating an in-page image editor or automating edits from the host. Not for desktop Photoshop plugins, raw postMessage wiring, or server-side image libraries (Sharp/Pillow); never call createEmbed on a zero-size container.
    1
    install
  92. Python Scala Collections · kayforkind
    Maps Python list, dict, and set operations onto Scala List, Map, and Set: typed empty creation, map/filter/foldLeft, lift/getOrElse, groupBy, zipWithIndex field order, and local ListBuffer snapshots. Use when the port's data-structure layer is the gap. Do not use for class/trait OOP, type-class or LazyList functional ports, or camelCase Pythonism cleanup.
    1
    install
  93. Search Company Knowledge · kayforkind
    Searches internal Confluence, Jira, and company docs (Rovo/Atlassian search) in parallel and cites what it finds. Use when looking up internal systems, terminology, deployment, or our docs. Not for public web search, academic papers (papers-skill), or questions about other companies.
    1
    install
  94. Suricata Offline Evejson · kayforkind bundle
    Runs Suricata offline against PCAPs (-r) and validates alerts in eve.json with jq — syntax check, positive/negative captures, rule regression. Use when testing Suricata rules, forensic PCAP review, or eve.json alert counts. Not for live IDS/IPS on a network interface, general jq tutorials (jq), or Suricata versions older than 6.
    1
    install
  95. Technical Diagram Expert · kayforkind
    Writes C4 Context through Deployment and Dynamic maps plus Mermaid flow, sequence, ERD, state, Gantt, and gitGraph into docs/architecture markdown. Trigger on system maps, API sequences, or sprint Gantt in PRs. Never embed diagrams as source comments; not a CSS/HTML implementation chair.
    1
    install
  96. Threejs Game Development · kayforkind bundle
    Owns the Three.js r160+ layer of a browser 3D game: scene graph, glTF with DRACO/KTX2, animation, Rapier or cannon-es, EffectComposer, dispose, and Vite. Use when starting or fixing a vanilla Three.js/WebGL web game (r3f only if asked). Not for the HTML5 game shell (browser-game-architecture), cross-engine FPS diagnosis (webgl-performance-tuning), Godot/Unity, or 2D Canvas/Pixi.
    1
    install
  97. Voice Activity Detection · kayforkind
    Returns (start_sec, end_sec) speech spans with Silero, SpeechBrain, or WebRTC VAD, then merge-and-pad postprocess. Use when gating diarization, ASR chunking, or push-to-talk so silence never reaches the next model. Do not use as a leading-silence energy trimmer (silence-detector) or as speaker-ID clustering (speaker-clustering).
    1
    install
  98. Web Interface Guidelines · kayforkind bundle
    Applies Vercel web interface guidelines: accessibility, CLS/visual stability, performance, forms, and reduced-motion. Use when reviewing or building UI markup, styles, or animation for a11y/performance/visual-stability compliance. Not for React/Next scaffolding or bundle scripts (senior-frontend). Not a motion-taste chair (emil-design-eng).
    1
    install
  99. Webgl Performance Tuning · kayforkind
    Profiles and tunes runtime FPS of canvas WebGL/WebGPU games via draw-call batching, instancing, KTX2, object pools, GC pauses, and engine knobs for Three.js, Phaser, PixiJS, Babylon.js, and Godot 4 Web. Use when FPS drops, jank, leaks, or context-lost appear. Not for native Unity/Unreal/Godot desktop builds, Core Web Vitals on document pages, or k6 API load tests.
    1
    install
  100. Apify Actor Development · kayforkind bundle
    Creates Apify Actors with the apify CLI, Crawlee Cheerio/Playwright crawlers, .actor schemas, local apify run, and apify push. Use when the user is bootstrapping or deploying an Actor on the Apify platform. Never confuse this with generic Playwright test scaffolding or a raw HTTP scraper outside Apify.
    1
    install