Skills — le catalogue de compétences de vos agents

Generate 100% original Farsight commercial video intros by analyzing structural elements of inspiration videos while strictly avoiding copyright infringement.

“手动介入式分片视频下载：用有头浏览器（用户可见、不自动关闭）打开视频页，拦截真实流地址，多路并行高速下载分片（TS 流 / m3u8 / 大分片 MP4），无损合并为完整 MP4，验收通过后再清理分片。面向需要分片下载的流媒体视频，尤其是需要登录、短信验证码、滑动验证、观看密码或切换清晰度的受限站点（优酷、腾讯视频、爱奇艺、B 站等）。关键词：下载视频、视频下载、优酷下载、需要登录的视频、需要密码的视频、m3u8、ts 分片合并”

Build and harden a skill with evals — interview to design its eval tasks, then run, measure, and iterate. Use when the user wants to create or improve a skill's eval, or run the create → test → improve loop for a skill.

Use when installing, repairing, reusing, inspecting, or verifying readiness of the native NVIDIA Video Codec SDK or PyNvVideoCodec on Jetson, including the one-frame encode/decode smoke test with official samples, and when interpreting what those readiness results, including CPU-buffer and device-memory sample modes, do and do not establish.

Use when planning, executing, and independently validating Jetson Video Codec SDK or PyNvVideoCodec encode/decode, transcode, segmentation, container decode, AV1, or concise acceptance workflows.

Use when turning a Jetson encoder use case into one surface-neutral recipe with native Video Codec SDK and PyNvVideoCodec projections.

从零创意生图，也可定向改模特、姿势、搭配。一句描述（可选参考图）→ 图。当用户说「创意生图」「生成一张」「改个姿势」「换模板」「随便来张图」时使用。

script to video, screenplay to video, shot list to video — turn a script into a storyboarded, shot-by-shot video: break down scenes, generate shots, assemble, validate. Use when the user gives a script or scene breakdown and wants a video.

Knowledge-course / online-teaching avatar lecturer — 16 landscape base-portrait recipes plus the slide-first pipeline that keeps a 20-minute lesson under 3000 credits. Use when the user wants a digital human to teach a paid course, a training module or a knowledge-sharing video on platforms like Xiaoetong, Tencent Classroom, Bilibili, YouTube or an internal LMS.

商品短视频广告。卖点 → 分镜脚本 → 分段生成 → 拼接加字幕成片。当用户说「做条广告」「短视频」「投流素材」「分镜脚本」「带货视频」时使用。

静态主图转主图视频。一张商品图 → 3–5 秒可上架的主图短视频。当用户说「主图视频」「图转视频」「让图动起来」「加个视频」时使用。

多单品融合成一整套 Look。最多 8 张单品图 → 同一模特身上的完整搭配商拍图，每件单品保真。当用户说「多件搭配」「融图」「组一套 look」「搭配图」「几件衣服合成一张」时使用。

1. Get the storyboard info

Image-to-SVG tool: converts raster images (PNG/JPG) into color vector SVG and returns the URL, suitable for lossless scaling and vectorization of logos, icons, and flat illustrations.

product video, product demo video, product ad video, product showcase — turn a product's photos or link into a polished demo / ad video. Use when the user wants a product demo, showcase, or ad video.

MiniMax Hailuo omni-modal video model with native stereo audio, producing 5-15 second clips at up to 2K. Supports text-to-video, first/last frame transitions and multi-asset references for character and scene consistency.

ecommerce video, product video, shopping ad video, ecommerce video generator — turn a product (photos or link) into a conversion-focused ecommerce ad video. Use when the user has a product and wants a store, TikTok, or ad video.

article to video, text to video, news to video, essay to video — turn a written article into a narrated explainer video: outline, storyboard, voiceover, build, validate. Use when the user pastes or gives an article and wants a video.

Use when authoring or validating PAIDF augmentation YAML configs, or running remote Cosmos Transfer (including Cosmos3 WSM controls), Cosmos Predict, image-edit, or image-to-video inference.

Read-only Ninebot/九号 electric vehicle status and report skill. Query battery, range, charging, smart-service remaining days, location, monthly/today rides, recent ride details, and generate map or ride-track images through a locally authenticated ninecli session. Use when the user asks about 九号/Nine

pcb; pcb modify/resume/reprice/package: design and verify editable KiCad boards, JLCPCB manufacturing bundles, and assembly pricing with supplier review.

Run the disk-backed DEFT AOI improvement loop for NVIDIA Cosmos Reason 3 / Cosmos3 models, using Nano by default and Edge or Super when explicitly requested: evaluate the base model on Proxy and frozen Benchmark splits, mine real image pairs from Proxy gaps, assemble a per-iteration Train JSON from selected Mining samples, train with cosmos-rl LoRA SFT, and repeat through the selected platform's submit/status/logs/cancel contract. This migration supports bare labels only: the assistant response must be exactly OK or NG. Use for "run Cosmos3 DEFT AOI", "CR3 AOI loop", or "improve Cosmos3 PCB in

The contract home for TAO's SDK-free execution pipeline — authoritative JSON Schemas for the four typed artifacts (spec-bundle, job-record, results_dir layout, best_rec) plus the fixed job-status vocabulary and the nested-not-dotted spec rule. Use when authoring or validating a spec-bundle before submit, writing or reading a .tao/jobs job-record, resolving where results land, or consuming AutoML's best_rec.json. Trigger phrases include "validate the spec bundle", "job record schema", "status vocabulary", "results_dir layout", "best_rec schema".

音频流上传专业版 —— 面向企业团队与专业创作者的高级音频上传工具。核心能力: - 批量音频上传，支持队列管理与断点续传 - 完全自定义编码配置：比特率、采样率、声道、编解码器 - 多质量预设输出（标准/良好/最高/无损），满足不同播放场景 - HLS与DASH双流媒体格式支持，适配多终端播放 - 丰富的元数据管...