Media Skills

HX Audio Player for music and sound playback. Use when implementing music playback, sound effects, HXMusic, HXSound, or audio in HXAudioPlayer. Library is in the hxaudio module.

Audit for performance and battery - layout thrash, heavy animations, timers, background tasks. Identifies hotspots affecting smoothness and power consumption. Use before releases or when users report lag/battery drain.

Design meaningful interactions, microinteractions, animations, state machines, gesture patterns, error prevention, and product behavior. Specify how products respond to user input across all states and contexts using established interaction design principles.

Create customer journey maps, service blueprints, experience maps, empathy maps, and other alignment diagrams. Guides diagram type selection, content structure, illustration syntax, and alignment workshops.

根据用户描述生成高质量绘图 prompt，并按通用、roadmap、schematic 模式调用 gpt-image-2 或 Nano Banana/Gemini 图片模型 API；gpt-image-2 默认使用低画质、原生尺寸和 JPEG，第 2 轮起基于上一轮图片做保真微调。

Phase-1 (S4) of a comic movie — PRODUCE the canonical reusable references the whole spiral conditions on. Per asset it bakes ONE canonical 1:1 white-bg identity ref via the agent mcp__codex__codex sidecar bake (Codex native image_gen — conditioned, never hand-pasted) OR, for a deterministic motif (clock/chart/stamp/star-map), emits a single-source parametric SVG from a python generator (asset_lib.py). Hashes + base64-encodes the bake into the asset node's output_ref (all 6 fields or it's a schema violation), versions every ref _v{NNN} with a supersedes self-edge, and runs a single-source collision gate. GENERATES ONLY — it NEVER self-locks/approves (that is the cross-model asset-review-loop's job, a different model family). Identity is BRING-YOUR-OWN; the ARIS chibi duo is only the worked example.

The authoring-side DIEGETIC continuity AUDIT — it produces the world-state evidence that comic-cross-layer-gate `--gate continuity` adjudicates against the storyboard's MOTIF STATE TABLE (motif_ledger). It checks each baked panel STRICTLY against the pre-committed table row, then emits a continuity review-node the gate fuses (it does NOT own a bake-time KEEP). Gemini auto-gemini-3 is the per-panel FACT extractor (one analyzeFile per panel image, fact-only); Claude is the deterministic JUDGE against the table row + the global invariants (DDL countdown monotonic never-rewind, bounce S02 = the film's ONLY MAX, the two metric columns never co-mingle, the non-adjacent MIRROR LOCKS). CAST-AWARE: a panel simply NOT containing a character is design, NEVER a drift penalty (absence≠drift). On a same-panel miss it requests `retry_panel`; a cross-frame motif break is routed into the page assembly_gate's drift set (seed-anchored comic panels are independent — no cross-panel rollback). Use when the user says "查连续性", "continuity audit", "连戏检查", "world-state drift", "motif table check", "对一下分镜表", or a panel has been baked and you need to verify it realized its motif-table row.

The ONE parameterized score-fuser for EVERY comic-author authoring gate — `--gate intent|outline|asset|storyboard|blueprint|continuity|p0_proof|compile`. A single fuser (not a per-layer split) prevents drift. It NEVER re-runs a reviewer; it collects the reviewer score-nodes already on the wiki (via `reviews` edges), fuses them deterministically (min-fuse per dim, max for inverted dims, SKIP missing dims — never substitute 0), then a Codex xhigh adjudicator (NO model pin — follows the local codex config) that sees ONLY structured inputs (scores + tags + raw artifact PATHS + verbatim source context + verbatim rubric — NEVER reviewer prose) makes an asymmetric call (threshold HARD-vetoes "advance"; Codex SOFT-vetoes everything else). The `--gate p0_proof` mode is the zero-credit pre-production proof: a text-only cross-model adversarial review of the pipeline's CODE + IR-CONTRACT + ENGINE state-machine that MUST clear all blockers in BOTH non-author families and then MINT the digest-bound decision:p0_proof certificate via scripts/run_p0_proof.py BEFORE a single metered image-generation credit is spent. Use when a sibling step (intent-parser, outline-creator, asset-review-loop, storyboard-creator, blueprint-author, continuity-audit, json-compiler) defers its acquittal to "the gate", or the user says "过 gate", "cross-layer gate", "审这一层", "p0 proof", "证明流水线再花钱".

Phase-1 Layer-0 of the comic-author suite — turn ANY raw idea / a locked video skeleton / an audience note into ONE schema-valid intent_spec node that fixes the logline (an editorial climax), the tagline, and the named DUAL-IDENTITY design constraint the whole comic optimizes against. Codex-as-parser (raw bytes + schema, never Claude's gloss) + reviewer-independence + banned-vocab lint + confidence-gated under_review routing + a HARD user-approval gate. It NEVER invents panels, assets, or storyboard detail, and NEVER silently defaults — every assumption becomes an uncertainties[] entry. Use it whenever a comic starts from a fuzzy brief and you need a locked, auditable premise before outlining.

Generate a publication-grade method / architecture / pipeline / workflow figure (a paper or README 'Figure 1') as an AUDITABLE object, not a one-shot prompt. A deterministic JSON blueprint LOCKS the content; an image model (gpt-image-2, baked by the agent via mcp__codex__codex — Codex GPT-5.5 xhigh, sandbox workspace-write) bakes the aesthetic from a labeled-condition render + the project's real identity refs; a cross-model panel (Gemini + Codex) blind-transcribes the result and a script hard-diffs it against the blueprint; the loop regenerates until Gemini AND Codex approve and the diff is empty — then the calling agent (Claude) gives the structural sign-off. NOT for statistical plots (use a plotting tool) or photo scenes.

> Image optimization analysis for SEO and performance. Checks alt text, file sizes, formats, responsive images, lazy loading, and CLS prevention. Use when user says "image optimization", "alt text", "image SEO", "image size", or "image audit".

Use when the user wants to repurpose a YouTube video for Bilibili, add bilingual (English-Chinese) subtitles to a video, or create hardcoded subtitle versions for Chinese platforms.

Add the canonical Clawd character to a user-provided photograph by selecting one bundled pose reference and compositing it with the photo through built-in image generation. Use this skill whenever the user asks to put, add, place, hide, hang, sit, stand, sleep, lean, climb, hug, push, pull, or otherwise make Clawd interact with an object in a photo, including requests phrased in Chinese such as “把 Clawd 放进图里”, “让它探头”, “挂在屏幕上”, or “抱住电线”.

>- Create original pixel-art "Clawd" crab emotes as pure SVG + CSS keyframe animations (no frame-by-frame GIFs, no raster sprites), and optionally export them to looping GIFs. Each emote is a tiny pixel crab doing an activity — holding a prop, wearing a hat, reacting — built from <rect>/<polygon>/<circle> and animated with @keyframes. Use this skill whenever the user wants to make, add, fix, or extend Clawd / pixel-crab / desk-pet emotes, mascot animations, "表情/图鉴" cards of an animated character, festival or daily-life crab stickers, or asks to turn such SVG animations into GIFs — even if they don't say the word "SVG". Also use it when fixing visual bugs in these emotes ("帽子悬空/hat floating", "脚断开/feet detached", a prop that looks disconnected from the hand), since the fixes follow specific structural rules documented here.

Transform audio recordings into professional Markdown documentation with intelligent summaries using LLM integration

Generates optimized prompts for any AI tool. Use when writing, fixing, improving, or adapting a prompt for LLM, Cursor, Midjourney, image AI, video AI, coding agents, or any other AI tool.

| 月老鳌烨的婚恋咨询互动体验。基于409个视频、69万字语料深度蒸馏， 提炼完整的连麦互动流程引擎、7条核心公理、6套方法论和精确的表达DNA。 用途：以鳌烨的身份与用户做连麦式婚恋咨询，完成从信息采集到方案推荐的完整闭环。 当用户提到「鳌烨」「月老」「连麦」「婚恋分析」「帮我看看条件」「打个分」时使用。 即使用户只是说「用鳌烨的方式聊聊」「月老模式」「鳌烨上线」也应触发。

Amazon 图片诊断助手适合运营、产品、销售、software在用户提出“主图能让人买吗”这类问题，需要快速拆解目标、判断重点并形成可执行结果时使用，帮助基于输入材料生成商品/店铺诊断、卖点与风险提示、运营动作建议。

视频质量诊断助手适合市场营销、运营、内容媒体、电商在用户提出“这段视频够发布吗”这类问题，需要快速拆解目标、判断重点并形成可执行结果时使用，帮助基于输入材料生成摘要、诊断结论、行动建议和可复用交付物。

图片压缩助手适合内容创作者、市场营销、运营、内容媒体在用户提出“图片能再小一点吗”这类问题，需要快速拆解目标、判断重点并形成可执行结果时使用，帮助基于输入材料生成视觉/创意诊断、提示词或分镜方案、发布规格检查。

封面图生成助手适合内容创作者、市场营销、运营、内容媒体在用户提出“封面能吸引点击吗”这类问题，需要快速拆解目标、判断重点并形成可执行结果时使用，帮助基于输入材料生成视觉/创意诊断、提示词或分镜方案、发布规格检查。

图片生成助手适合内容创作者、市场营销、运营、内容媒体在用户提出“这张图能用吗”这类问题，需要快速拆解目标、判断重点并形成可执行结果时使用，帮助基于输入材料生成视觉/创意诊断、提示词或分镜方案、发布规格检查。

图文卡片生成助手适合内容创作者、市场营销、运营、内容媒体在用户提出“卡片图好读吗”这类问题，需要快速拆解目标、判断重点并形成可执行结果时使用，帮助基于输入材料生成视觉/创意诊断、提示词或分镜方案、发布规格检查。

视觉质量诊断助手适合市场营销、运营、内容媒体、电商在用户提出“这张图够发布吗”这类问题，需要快速拆解目标、判断重点并形成可执行结果时使用，帮助基于输入材料生成摘要、诊断结论、行动建议和可复用交付物。

视觉质量诊断助手适合内容创作者、市场营销、运营、内容媒体在用户提出“这张图够发布吗”这类问题，需要快速拆解目标、判断重点并形成可执行结果时使用，帮助基于输入材料生成创意方案、提示词/脚本、制作清单。

视觉质量诊断助手适合市场营销、运营、内容媒体、电商在用户提出“这张图够发布吗”这类问题，需要快速拆解目标、判断重点并形成可执行结果时使用，帮助基于输入材料生成摘要、诊断结论、行动建议和可复用交付物。

视觉质量诊断助手适合销售、运营、市场营销、电商在用户提出“这张图够发布吗”这类问题，需要快速拆解目标、判断重点并形成可执行结果时使用，帮助基于输入材料生成视觉/创意诊断、提示词或分镜方案、发布规格检查。

视觉质量诊断助手适合市场营销、运营、内容媒体、电商在用户提出“这张图够发布吗”这类问题，需要快速拆解目标、判断重点并形成可执行结果时使用，帮助基于输入材料生成摘要、诊断结论、行动建议和可复用交付物。

Remotion 实践助手适合technical、运营、software、内容媒体在用户提出“视频代码够稳吗”这类问题，需要快速拆解目标、判断重点并形成可执行结果时使用，帮助基于输入材料生成摘要、诊断结论、行动建议和可复用交付物。

AI 视频剪辑诊断助手适合市场营销、运营、software、内容媒体在用户提出“这段素材怎么剪”这类问题，需要快速拆解目标、判断重点并形成可执行结果时使用，帮助基于输入材料生成剪辑结构诊断、片段取舍和时间点建议、字幕配音与平台重构清单。

视频质量诊断助手适合市场营销、运营、内容媒体、电商在用户提出“这段视频够发布吗”这类问题，需要快速拆解目标、判断重点并形成可执行结果时使用，帮助基于输入材料生成摘要、诊断结论、行动建议和可复用交付物。

视频转 GIF 优化助手适合市场营销、产品、technical、software在用户提出“这段视频怎么转 GIF”这类问题，需要快速拆解目标、判断重点并形成可执行结果时使用，帮助基于输入材料生成GIF 转换参数建议、FFmpeg 转换命…

AI 视频生成策划助手适合市场营销、software、内容媒体、电商在用户提出“想生成什么视频”这类问题，需要快速拆解目标、判断重点并形成可执行结果时使用，帮助基于输入材料生成视觉/创意诊断、提示词或分镜方案、发布规格检查。

智能配图助手适合内容创作者、运营、technical、内容媒体在用户提出“还要配图，好麻烦”这类问题，需要快速拆解目标、判断重点并形成可执行结果时使用，帮助基于输入材料生成摘要、诊断结论、行动建议和可复用交付物。

B站短视频运营增长助手适合内容创作者、运营、品牌方、电商在用户提供了 B 站视频链接时使用，帮助基于输入材料生成评论情绪和讨论结构、观众画像和兴趣信号、内容与互动建议。

业务诊断助手适合市场营销、运营、service、销售在用户提出“这件事该怎么做”这类问题，需要快速拆解目标、判断重点并形成可执行结果时使用，帮助基于输入材料生成摘要、诊断结论、行动建议和可复用交付物。

抖音短视频运营增长助手适合内容创作者、运营、品牌方、电商在用户提供了抖音内容链接时使用，帮助基于输入材料生成情绪和舆情判断、用户画像和意图信号、运营建议和回复建议。

视频质量诊断助手适合市场营销、运营、内容媒体、电商在用户提出“这段视频够发布吗”这类问题，需要快速拆解目标、判断重点并形成可执行结果时使用，帮助基于输入材料生成摘要、诊断结论、行动建议和可复用交付物。

视觉质量诊断助手适合市场营销、运营、内容媒体、电商在用户提出“这张图够发布吗”这类问题，需要快速拆解目标、判断重点并形成可执行结果时使用，帮助基于输入材料生成摘要、诊断结论、行动建议和可复用交付物。

快手短视频运营增长助手适合内容创作者、运营、品牌方、电商在用户提供了快手内容链接时使用，帮助基于输入材料生成情绪、画像和讨论焦点、舆情与转化线索、优化建议。

视频质量诊断助手适合内容创作者、运营、市场营销、设计在用户提出“这段视频够发布吗”这类问题，需要快速拆解目标、判断重点并形成可执行结果时使用，帮助基于输入材料生成摘要、诊断结论、行动建议和可复用交付物。

小红书短视频运营增长助手适合内容创作者、运营、品牌方、电商在用户提供了小红书笔记链接时使用，帮助基于输入材料生成情绪和舆情视图、用户画像和意图分析、优化与转化建议。

>- Generate a hand-drawn (sketch-style) architecture or flow diagram as a committable image (SVG + PNG). Use when the user wants a sketch-aesthetic diagram for a README, docs, or hackathon/Devpost submission that still has source and renders it with mermaid-cli — works on any viewer and on Devpost. '손그림 아키텍처', '스케치 도식', 'handDrawn', 'Devpost diagram', '손그림으로 그려줘', '아키텍처 손그림', 'sketch-style architecture'.

Generate AI images using OpenAI's gpt-image-1 model with customizable aspect ratios and artistic themes. Use when the user wants to create images, generate artwork, or mentions image generation with specific styles like Ghibli, futuristic, Pixar, oil painting, or Chinese painting.

GSAP animations for JARVIS HUD transitions and effects

Expert skill for implementing speech-to-text with Faster Whisper. Covers audio processing, transcription optimization, privacy protection, and secure handling of voice data for JARVIS voice assistant.

Expert skill for implementing text-to-speech with Kokoro TTS. Covers voice synthesis, audio generation, performance optimization, and secure handling of generated audio for JARVIS voice assistant.

Expert skill for implementing wake word detection with openWakeWord. Covers audio monitoring, keyword spotting, privacy protection, and efficient always-listening systems for JARVIS voice assistant.