claude-code-config

Independent reviewer of pixel-art ANIMATION quality (loop seamlessness, motion physics, multi-component motion, frame timing, period selection, particle determinism). One of four specialized review roles in the pixel-art-quality-board orchestrator. Use when the user asks to "check animation timing", "verify loop seamlessness", "review motion quality", "is this animation organic", or proactively after pixel-art-storyboard produces an animated cover. Does NOT review static palette, surface detail, or composition silhouette — only animation behavior over time. Returns JSON verdict with scores per dimension (loop seamlessness, motion physics, multi-component, particle determinism, period appropriateness) and specific fixes.

Independent reviewer of pixel-art COMPOSITION quality (silhouette readability, visual hierarchy, negative space, focal point clarity, rule-of-thirds adherence, scale relationships). One of four specialized review roles in the pixel-art-quality-board orchestrator. Use when the user asks to "check composition", "is this readable as a thumbnail", "verify visual hierarchy", "review focal point", or proactively after pixel-art-storyboard produces output. Does NOT review palette, surface detail, or animation timing — only spatial composition and silhouette. Returns JSON verdict with scores per dimension (silhouette test, focal point, hierarchy, negative space, scale, framing) and specific fixes.

Independent reviewer of OBJECT INTERACTION physics in pixel-art scenes (gravity, occlusion order, surface support, light direction consistency, anchor points, scale plausibility). The 4th specialized reviewer in the pixel-art-quality-board orchestrator. Use when the user asks "do objects interact correctly", "is anything floating", "check physical plausibility", "verify light direction consistency", or proactively after pixel-art-storyboard produces output. Catches things that style/animation/composition reviewers miss - e.g. chess pieces floating ABOVE the board, character shadow falling RIGHT while highlight is on RIGHT (light from same direction as shadow), object in foreground with no occlusion of background it should overlap. Returns scored JSON verdict with per-axis findings + specific fixes.

Orchestrator agent that runs 4 specialized pixel-art reviewers in parallel (style + animation + composition + interaction), then synthesizes their verdicts into a single PASS/NEEDS_WORK/REJECT decision with prioritized fixes. Use when the user asks to "review my pixel art quality", "score this animation", "comprehensive pixel art audit", "is this retouch-quality", or proactively after pixel-art-storyboard / pixel-art-studio produce output. This is the standard quality-control entry point. The interaction reviewer catches things that style/animation/composition miss — e.g. chess pieces floating ABOVE the board, character shadow falling from wrong direction, pieces with no surface support. Returns synthesis JSON with per-reviewer verdicts, dimension scores, blocking issues, and ranked fix priority.

Use when the user generated pixel art and wants an independent quality review, or when the pixel-art-studio skill has produced output that needs to be evaluated before delivering. Performs Generator-Evaluator review: runs quality_check.py, reads the rendered PNG with fresh context (does not see how the art was made), and writes PASS / HOLD / REJECT verdict with specific findings. Trigger on: "review my pixel art", "score this sprite", "is this pixel art good", "check sprite quality", "пересмотри пиксель арт", "评估像素画质量", or proactively after the pixel-art-studio skill produces a PNG/JSON.

Designing agent harnesses and tool systems — risk taxonomy for tools, permission decisions, draft/commit pattern, structured tool results, agent budgets (10 types), context trust labels against prompt injection, plan-artifact, approval records, observability and traces, evals (13 categories), event model, streaming buffering, 3rd-party skill install checklist. Use when building a new Agent SDK app, custom orchestrator, MCP server, or Cloudflare Worker with tool calls; when designing tools and permissions; when writing an agent loop; or when you need trust labels for external content.

On-demand minimalism intensifier — write the leanest correct code, kill over-engineering before it starts. Use when the user says "be lazy / yagni / simplest / minimal / shortest / don't over-engineer", complains about bloat / boilerplate / unnecessary dependencies / abstraction, or before writing a substantial chunk of new code. Supports intensity: lite / full / ultra. Pairs with the always-on quality-code rule and the over-engineering-advisor hook.

Практическая инженерия диффузионных моделей: архитектуры, обучение, инференс, оптимизация памяти. Использовать при любых задачах с диффузионными моделями: проектирование или модификация архитектуры (UNet/DiT/Flow/Flux), выбор и настройка schedulers/samplers, дообучение (LoRA/DreamBooth/full fine-tune), оптимизация памяти (AMP/checkpointing/ZeRO/FSDP/quantization), замена или fusion текст-энкодеров (CLIP/Qwen), работа с Diffusers, отладка диффузионных пайплайнов, оценка качества (FID/CLIPScore/LPIPS), latent diffusion, VAE, guidance/CFG, rectified flow, Stable Diffusion, SDXL, Flux. Также применять при вопросах про GPU-память при обучении генеративных моделей, text-to-image пайплайны, ControlNet, multi-encoder fusion, WebDataset.

Expert prompt engineering for FLUX.2 [klein] image generation and editing model. Use this skill whenever the user wants to create prompts for FLUX.2 [klein], generate images, edit photos with the klein model, work with multi-reference image editing, or needs templates for T2I/I2I tasks. Trigger for any mention of: FLUX.2, flux klein, BFL API, image editing prompts, text-to-image prompts for FLUX, product mockups, poster generation, UI mockups, sticker packs, character design, seamless textures, or any request to write/improve/translate prompts for FLUX-family models. Also trigger when user asks about guidance_scale, inference steps, distilled vs base modes, or multi-reference workflows.

Comprehensive reference for training LoRAs on FLUX.2 Klein 9B and Qwen Image Edit 2511 models. Use this skill whenever the user asks about: training LoRAs for flux2/flux 2 klein/qwen-image-edit, before/after edit LoRAs (head swap, face swap, image editing), inpainting LoRAs, training at larger resolutions, latent space expansion, VAE fine-tuning, multi-reference training (2 input images → 1 output), dataset preparation for edit models, zero_cond_t, ai-toolkit/SimpleTuner/DiffSynth configs, BFS head swap LoRA methodology, Qwen Edit architecture, consistency mode, dual encoding, FuseAnyPart, ACE++, maximum training resolution, или любые вопросы об обучении диффузионных моделей. ВСЕГДА используй этот скилл.