madebyaris

Build distinctive, performant, production-grade 3D on the web with Three.js, React Three Fiber, and WebGL. Use when the user asks to "build a 3D scene", "add a 3D hero/landing", "make a product viewer/configurator", mentions three.js / react-three-fiber / r3f / drei / webgl / shaders, or asks to make an existing 3D scene "look good", run smoothly, or work on mobile. Covers aesthetic direction (anti-slop for 3D), lighting/material/post-processing, performance budgets, responsive WebGL, graceful degradation, accessibility, and M3 multimodal parity from reference video/frames.

Category-aware design skill that builds distinctive, production-grade UIs. Brand-vs-product register, color strategy commitment, scene-based theme choice, palettes, font pairings, UX patterns, shadcn/token integration, empty-error-loading copy, secondary slop signals, multimodal design parity from mocks, and a pre-ship second pass. Framework-agnostic.

Conducts multi-step deep research on any topic using iterative search, reflection, and synthesis. Use when the user asks to research, investigate, survey, compare, analyze, deep-dive, or explore a topic in depth. Covers web research, codebase analysis, documentation review, mixed-source investigation, and M3 long-context compression discipline.

Production-style incident triage workflow for logs, metrics, code, safe mitigations, and M3 multimodal visual evidence (screenshots, screen recordings). Use when debugging alerts, regressions, outages, or suspicious runtime behavior.

How to use MiniMax M3's 1M-token MSA context productively: what to load vs. compress, when to retrieve vs. ingest, how to keep skills shallow in the always-on prompt and deep in skills, and how to plan retention across iterations. Load when the task might exceed ~200K tokens, when the user asks to "keep all of this in mind", or when you are tempted to start a fresh session to "free context".

How to use MiniMax M3's native multimodal input (image, video) for grounded decisions in coding work. Covers reading attached images/frames, treating them as ground truth for visual claims, screenshot diffing, design parity from mockups, and routing visual evidence through reports and PRs. Load when the user attaches an image, screenshot, mockup, frame, or short clip — or when the task involves "make it look like this", "match this design", "why does this UI look wrong", or "read this error screenshot".

MiniMax-native multimodal workflow for image, video, voice, music, and media-processing tasks. Use when the user asks to generate image/video/audio assets, wants MiniMax-specific media APIs, needs TTS or voice workflows, wants reproducible local media outputs, or needs FFmpeg-style processing around generated media. M3's native multimodal input means image/video inputs can be fed directly to the model for grounded decisions in coding work.

Execute planned implementations following todo-lists systematically. Use for code generation, building features, and executing SDD plans.

Recall and persist durable project knowledge (decisions, conventions, gotchas, architecture) across SDD sessions. Provider-aware — supports standard (rules-only), Cursor native Memories, and mem0. Use at the start of planning/implementation to recall context, and at the end to persist discoveries.

Update specifications with discoveries made during development. Use when implementation reveals new requirements, constraints, or design changes.

Generate technical plans from specifications. Use when creating architecture documents, designing system components, or preparing for implementation.

Pattern investigation and technical research before specification. Use when technical approach is unclear, exploring existing solutions, or analyzing codebase patterns. Supports deep research mode for thorough external investigation.

Compare implementation against specifications, identify gaps and issues. Use for code review, quality assurance, and verifying spec compliance.

Spawn effective OpenClaw sub-agents — clear briefs, isolated vs fork context, no polling loops, cheaper models for workers.

End-to-end software work—read codebase, implement, test, and summarize changes like a senior engineer.

UX/UI feedback and copy tuned for clarity, accessibility, and consistency—without redesigning everything unprompted.

Lawful, ethical open-source intelligence on public data only—no stalking, no bypassing access controls, no invasive targeting.

Rigorous public research using web search, fetch, and source triangulation. Use for technical docs, competitors, and fact-checking.

Guides the agent to build distinctive, production-grade UIs avoiding generic "AI slop" aesthetics. Enforces a pre-coding design thinking phase, bold aesthetic variety, strict hierarchy, real icons/imagery, responsive design across desktop/mobile. Framework-agnostic.

Conducts multi-step deep research on any topic using iterative search, reflection, and synthesis. Use when the user asks to research, investigate, survey, compare, analyze, deep-dive, or explore a topic in depth. Covers web research, codebase analysis, documentation review, and mixed-source investigation.

Execute planned implementations following todo-lists systematically. Use for code generation, building features, and executing SDD plans.

Iterative multi-source investigation with contradiction logging—use for audits, vendor comparisons, security posture reviews, benchmarking methodology, literature surveys, ambiguous technical history, regulated domains, unfamiliar APIs, architectural landscape scans, falsifiable claims needing citations, stakeholder briefings where evidence tiers matter more than narration.

Version-pinned coding and framework best practices from official docs and corroborated senior/maintainer guidance — use for unfamiliar stacks, upgrades, migrations, architecture reviews, "best practice" questions, or before introducing a new pattern when style governance says improve might be warranted.