PRTV (https://prtv.pro) is a cloud digital signage editor: menu boards, promo screens and info channels for Smart TVs, set-top boxes and browsers, with live widgets (clocks, weather, rates, QR, countdowns, calendars, social feeds) built on https://s.prtv.su. This repository holds open Agent Skills for the PRTV editor: what an AI agent needs to know to build and edit slides in prtv.pro the way a person does, in English and Russian.
По-русски. PRTV (prtv.pro) — конструктор цифровых вывесок: меню-борды, промо-экраны и информационные каналы для Smart TV, приставок и браузеров; информеры (часы, погода, курсы, QR, обратный отсчёт, календари, соцленты) собираются на s.prtv.su. В этом репозитории — открытые скиллы для ИИ-агентов (Claude, Codex, Gemini CLI, Cursor и других): как собирать и править слайды в конструкторе prtv.pro через интерфейс, без внутреннего API. Каждый скилл — папка с SKILL.md (английский) и SKILL.ru.md (русский). Начинать с prtv-editor-overview. Лицензия CC-BY-4.0.
Open, reusable skills that teach an AI agent (Claude, Codex, Gemini CLI, Cursor or any tool that reads the Agent Skills format) how to work in the PRTV digital-signage editor — https://prtv.pro — the way a person does: through the interface, reading the page state when needed, never through private APIs.
Each skill is a folder with SKILL.md (English) and SKILL.ru.md (Russian). The Russian file is not a byte-for-byte translation but the same material written naturally.
How to use
skills/ (or .claude/skills/) directory — the YAML header tells the agent when to load each one. Start with prtv-editor-overview.SKILL.md (or SKILL.ru.md) into the conversation, or give the agent the raw file URL.git clone https://github.com/prtvbiz-design/prtv-skills.gitLicense: CC-BY-4.0. Product: PRTV (prtv.pro), cloud editor for menu boards, promo screens and info channels on Smart TVs, set-top boxes and browsers. Everything in these files was verified in the browser on real builds; where something was not verified, the text says so.
Load prtv-editor-overview first; the others assume it.
| Skill | What it covers | Load when |
|---|---|---|
| prtv-editor-overview | Slideshow → slide → element, the 1920×1080 space, URLs, panels, DOM conventions, fonts, uploads, TV slots, watermark | any task in the editor |
| prtv-slide-settings | Slide strip and its hover buttons, reordering, duration, 15 transitions, progress bar, background (colour / image / video), element defaults, templates, «Apply to all» | adding, configuring or checking a slide |
| prtv-element-layout | Position and size by dragging, rotation, depth, keep-aspect, permanent elements, the 7-item right-click menu, 57 entrance presets, clickable links for kiosks, layout rules (safe zone, type sizes, contrast, motion) | moving, layering, animating, linking an element; checking composition |
| prtv-text-element | Native text through TinyMCE: entering edit mode, toolbar, what persists and what is lost, the triple-click trap on menu rows, Ctrl+A replacement, quirks | writing or restyling text; text edits disappearing |
| prtv-html-block-animation | HTML block, CSS keyframes, SVG filters and SMIL without JavaScript, layers, ticker, recipes (banner behind a plane, TV static, wind, drift) | steam, shine, wind, looping effects; animating native text |
| prtv-video | Video element (YouTube, VK, Kinescope, Vimeo, own files), background video, embeds through HTML, sound on TVs, seamless-loop requirements | putting video on a slide or as a background |
| prtv-widgets | Widgets (informers): two ways to add, the three universal URL rules, parameter naming, two-level transparency, sizing fluid widgets, dependency risks, acceptance | any clock, weather, calendar, QR, countdown, finance, promo, map or social widget |
| prtv-widgets-catalog | Per-widget reference: builders, files, views, parameters, limitations, embeds — clocks, weather, countdown, QR, calendars, rates, promo card, moon, holidays, work-day, maps, TV programme, radio, timetable, poll, cloud feed, Yandex widgets, social feeds, RSS | building or reading a specific widget URL |
| prtv-agent-rules | Zero screenshots, computed coordinates, edits only through the interface, verification by reload and preview, JavaScript-tool traps, hover-only controls, slide activation, element timing, session pacing | the start of any automated session |
Every SKILL.md starts with a YAML header in the Agent Skills format:
---
name: prtv-<name>
description: what it covers and when to load it
license: CC-BY-4.0
metadata:
product: PRTV (prtv.pro) digital signage editor
version: "1.0"
date: "YYYY-MM-DD"
language: en | ru
scope: editor UI + reading page state from DOM; no private API calls
---
The body is a short introduction, numbered sections, a “What not to do” list and a “Related” line naming the neighbouring skills.
In: everything visible and doable through the editor interface; reading the page state (element ids, geometry, depth, active slide) with JavaScript; the public widget builders on s.prtv.su and their parameters; emulating input into interface fields.
Out: the editor’s private server methods, internal data-field names not visible in the interface, identifiers of any real project, and how to generate images or video with AI. A skill about producing menu-board graphics with AI is a separate publication.
https://raw.githubusercontent.com/prtvbiz-design/prtv-skills/main/<skill>/SKILL.md