// 📡 Pulse — Twice-daily tech briefing for group distribution. Designed to be triggered by cronjob at morning and evening, or manually. Use this skill only when the user says "pulse" or "/pulse".
Relay the current Claude Code session to Agentara for continued execution
Generate storyboard image boards and matching video-generation prompt scripts for specific scenes in a short video plan. Use this skill whenever the user asks to create a storyboard, storyboard image, video prompt script, scene prompt, image-to-video prompt, shot board, or per-scene video-generation package, especially when they specify a scene number, duration, or an existing video plan. This skill saves outputs as storyboard/scene-XX.png and storyboard/scene-XX.md and enforces grid sizing, timing labels, and strict character, wardrobe, prop, and location continuity.
Create cinematic movie poster concepts and final poster images from a user's brief, existing video plan, storyboard, character design, or project context. Use this skill whenever the user asks to generate a film poster, movie poster, key art, campaign poster, teaser poster, title poster, or promotional visual for a video project. This skill proposes several creative directions first when appropriate, defaults to a 3:4 poster ratio when no ratio is specified, and ensures typography, title treatment, imagery, genre, setting, and story tone are coherent and visually inventive.
Plan short-form video, vlog, travel, lifestyle, documentary, or cinematic social video scripts by negotiating duration, choosing a story arc and visual style, then splitting the film into timed scenes of 4-15 seconds. Use this skill whenever the user asks for a video plan, short video plan, scene breakdown, montage plan, travel video script, TikTok/Reels/Shorts storyboard, or cinematic short-form narrative, even if they only describe a loose theme or destination.
Create character design documentation and character design sheet images for video, storyboard, advertising, animation, or AI video-generation workflows. Use this skill whenever the user asks to design a character, extract a character from a reference image, make a character sheet, create a turnaround sheet, keep a person consistent across scenes, or generate character assets for a video project. This skill first writes a confirmable {character-name}.md design spec, waits for user approval or revision, and only then generates {character-name}.png.
Generate a beautifully designed infographic poster from an article URL or text content. Trigger when user says "article poster", "文章海报", "infographic", "信息图", "make a poster", "生成海报", "visual summary", or requests to convert an article/blog into a shareable image. NOT for generic poster design.
| name | pulse |
| description | 📡 Pulse — Twice-daily tech briefing for group distribution. Designed to be triggered by cronjob at morning and evening, or manually. Use this skill only when the user says "pulse" or "/pulse". |
A comprehensive twice-daily tech briefing combining product launches, open-source trends, curated news, market indices, and weather forecasts. Designed for group distribution to a 4000+ person tech community. Runs at morning and evening, or manually.
The final output should be in the user's preferred language which is zh-CN simplified Chinese.
Execute the following steps in order. Do NOT call web_fetch/web_search for sources that prefetch returned successfully — use the prefetch JSON when building Step 4. Step 2 is the only routine step that uses web_search (supplemental corporate and AI news). If a prefetch source’s errors entry is non-null, fall back to web_fetch/web_search for that source only.
cd .claude/skills/pulse && uv run scripts/prefetch.py
Returns a JSON blob with producthunt, github_trending, google_news, podcasts, weather, stock, and when GITHUB_OAUTH_TOKEN is set: agentara_stars, helixent_stars (GitHub API counts for MagicCube/agentara and MagicCube/helixent).
web_fetch/web_search (same coverage prefetch would have provided).Podcasts come from prefetch only — do not run a separate podcast search here.
Merge google_news from prefetch with supplemental items from the web searches below. Then apply curation rules (2c).
prefetch.searches, look for: earnings reports, major product launches, leadership changes, regulatory actions, acquisitions, layoffs, stock-moving events.prefetch.searches, look for: major model releases, notable funding rounds, regulatory developments, breakthrough research, significant product launches in the AI space.Before finalizing the news list, apply these filters strictly:
After Step 2, before assembling the final output:
CRITICAL — OUTPUT STARTS WITH THE TITLE, NOTHING ELSE. Your very first character of output MUST be
#. No preamble, no status updates, no "数据已收集完毕", no "正在整理", no "以下是今日Pulse", no transition sentences of any kind. The Pulse title IS the start of your response.
Use the following template. All section titles and product/repo/news names MUST be hyperlinked directly — no separate "sources" section.
Language: Chinese (full-width punctuation: ，、：！？。）for prose; English for product names, repo names, and technical terms.
# 📡 Pulse | {🌅 or 🌆} — {YYYY年M月D日}
## <font color="navy">🚀 Product Hunts</font>
- **[Product Name](https://www.producthunt.com/posts/slug)** {upvotes if available}
简短介绍与点评（简体中文）。
- **[Product Name](URL)**
...
(5-10 items)
---
## <font color="navy">🔥 GitHub Trending</font>
- **[owner/repo](https://github.com/owner/repo)** ⭐ {total stars} (+{today})
Description。{Language}。简短点评（简体中文）。
- **[owner/repo](URL)** ⭐ ...
...
(5 items)
---
## <font color="navy">📰 News</font>
- **[Headline](URL)**
1–2 sentence summary（简体中文）。
- **[Headline](URL)**
...
(3–8 items, or fewer if it's a quiet day)
---
## <font color="navy">🎙️ Podcasts</font> (If exists updates)
- **[Channel Name - Episode Title](episode_url)** — Podcast Name
shownotes 摘要（1-2 句简体中文重点总结）。
- **[Episode Title](episode_url)** — Podcast Name
...
(1-3 items)
---
## <font color="navy">💰 Stock Market</font>
> For each of the 4 indices in `prefetch.stock`, render the following block:
**{index_name} · {market}**
- 最新价：<font color='red/green'>**{price}**</font>
- 涨跌：{emoji} <font color='red/green'>**{chg:+.2f} / {pct:+.2f}%**</font>
- 最新交易日：{date}
> Color red if change is positive, green if negative.
> `{chart}` must be the `chart` field from that index’s object in `prefetch.stock`. Include the chart for every index.
{if_anomaly}
⚠️ {abnormal_description}
- {news_link_list}
{/if_anomaly}
> Repeat for all 4 indices: 上证指数, 深证成指, 纳斯达克综合, 道琼斯工业
---
## <font color="navy">🌤️ Weather</font>
**🧱 北京**
- 今天：{emoji} {type}，{low}°C ~ {high}°C
- 明天：{emoji} {type}，{low}°C ~ {high}°C
**🏙️ 上海（徐汇）**
- 今天：{emoji} {type}，{low}°C ~ {high}°C
- 明天：{emoji} {type}，{low}°C ~ {high}°C
**🗼 广州**
- 今天：{emoji} {type}，{low}°C ~ {high}°C
- 明天：{emoji} {type}，{low}°C ~ {high}°C
**🌴 深圳**
- 今天：{emoji} {type}，{low}°C ~ {high}°C
- 明天：{emoji} {type}，{low}°C ~ {high}°C
**🌊 杭州**
- 今天：{emoji} {type}，{low}°C ~ {high}°C
- 明天：{emoji} {type}，{low}°C ~ {high}°C
**🌳 南京**
- 今天：{emoji} {type}，{low}°C ~ {high}°C
- 明天：{emoji} {type}，{low}°C ~ {high}°C
{综合六座城市天气，带雨伞、洗车等建议}
---
> ✨ 本 Pulse 由 [Agentara](https://github.com/MagicCube/agentara) 智能生成{if agentara_stars} | 已有 {agentara_stars} 颗 ⭐ {/if}
> 📡 [打造你的专属 Pulse](https://github.com/MagicCube/agentara)，别忘了 ⭐ Star
> 👨🏻💻 [Helixent](https://github.com/MagicCube/helixent)开源 Coding Agent {if helixent_stars}已有 {helixent_stars} 颗 ⭐ ，欢迎试用{/if}
> **Footer rule**: If `prefetch.agentara_stars` is a number, show the star count after "智能生成". If `prefetch.helixent_stars` is a number, append the Helixent link and count as in the template. If either value is `null`, omit that repo’s star segment — do NOT fallback to web search or web fetch.
# 📡 Pulse | .... The # character must be literally the first character you output. Do NOT output any text before the title — no "数据已收集完毕", no "正在组装", no "以下是", no introductory sentences whatsoever. Violation of this rule is a critical formatting error.# for top-level, ## for sections, and list for individual items. This maps well to Feishu / Word export.- **[Title](URL)**. Do not put URLs only in a footnote or separate "sources" block.chart field on each entry in prefetch.stock as the image path in ![]() syntax (see template above).