๐ฏSkills33
Translates SRT subtitle files between languages with punctuation-bounded re-segmentation so cues end at real sentence breaks, supporting Simplified Chinese and English as first-class targets with bilingual SRT output.
Burns SRT subtitles into video using ffmpeg/libass for platforms that ignore embedded tracks (WeChat, Douyin), supports soft-mux togglable tracks, and handles full localization compositing (subs + dub + original audio bed) in a single encode pass.
A skill that transcribes audio/video into timestamped SRT subtitles, routing between Volcano ASR for Chinese and OpenAI Whisper for other languages, with word-level timestamps and punctuation-bounded cues.
A skill that generates time-aligned TTS voice dubs for videos from translated SRT subtitles, routing between Volcano TTS for Chinese and Edge TTS for other languages with optional multi-speaker support.
A post-production skill for video clips that adds AI-generated cover frames, HTML/CSS-quality captions synced to SRT, kinetic illustration overlays, chapter chips, and end-card CTAs โ all composed in a single HyperFrames project and rendered in one final encode. Typically used downstream of the wjs-segmenting-video skill to produce upload-ready MP4s.
Turns a single English word into a self-contained HLS video lesson by stitching together every clip where the word is spoken from the mira video base, prepending a bilingual word-intro card with IPA transcription, Chinese gloss, and TTS audio, and appending a CTA card.
Converts video between horizontal and vertical orientations (e.g., 16:9 to 9:16) by cropping a narrow band that tracks the active speaker using MediaPipe face landmarks and mouth-aspect-ratio variance, keeping the talking person in frame for platforms like TikTok, YouTube Shorts, and WeChat Channels.
Assists with writing and publishing WeChat Official Account (ๅพฎไฟกๅ ฌไผๅท) articles by lightly polishing drafts while preserving the author's voice, auto-generating cover images and explanation illustrations, and outputting content ready to paste into the WeChat backend.
Orchestrates the end-to-end video localization pipeline by routing through four sub-skills: audio transcription, subtitle translation, dubbing, and subtitle burning into the final video.
Segments long-form videos with SRT transcripts into 3 to 6 stand-alone topical short clips by analyzing the transcript, selecting topic boundaries, and cropping for the target platform.
Combines multiple synced camera angles into a single MP4 using audio-energy-driven camera switching, with optional picture-in-picture inset. Requires synced .sync.json sidecars from wjs-syncing-multicam.
A 5-step interactive reflection skill that guides users through personal setback analysis, moving from emotional review to behavioral training by identifying automatic responses, underlying patterns, new parameters to train, and concrete replacement actions. Based on the L3 "default response weight" framework for changing instinctive reactions rather than just intellectual understanding.
Automates daily promotion of Claude Code skills by researching marketplace trends, generating per-skill marketing plans, auto-posting to X (Twitter), and drafting community posts for Reddit, HN, and Discord. Includes idempotent scheduling with a 7-day cooldown per skill.
Computes time offsets for multi-camera recordings of the same event using audio cross-correlation and outputs lightweight .sync.json sidecar files without re-encoding originals. Uses log-energy envelope analysis at 8 kHz with multi-probe drift detection for robust synchronization across different microphones and recording devices.
A skill that converts Chinese WeChat articles (article.md) into narrated 1080x1920 portrait short videos (30-90 seconds) with TTS voiceover via Volcano TTS, HyperFrames CSS/GSAP animation, abstract watercolor backgrounds, and sound effects, outputting MP4 files for platforms like Douyin, Xiaohongshu, and Reels.
A skill for uploading videos to YouTube with support for batch uploads from a directory, per-video metadata via UPLOAD_META.md, and proxy-aware resumable uploads that bypass the stock Google API client connectivity issues. Designed for workflows behind SOCKS/HTTP proxies with 8 MB chunk uploads and exponential-backoff retry.
A roleplay skill that adopts Wang Jianshuo's thinking framework and writing style, distilled from deep research of roughly 1 million words of his English blog and 1.09 million characters of his Chinese blog spanning 2002 to 2022.
A holistic project-state audit skill that checks for unmerged branches, stalled PRs, failed CI runs, unreleased commits, and plan drift. It runs a read-only investigation first, presents a grouped checklist, and only applies fixes after explicit user confirmation.
A Claude Code skill that drafts and posts daily X/Twitter tweets by extracting the most quotable content from recently published WeChat public account articles, with automatic history tracking to avoid duplicates.
A skill that auto-syndicates WeChat public account (ๅพฎไฟกๅ ฌไผๅท) articles across social platforms including X, Bluesky, Threads, LinkedIn, and Facebook, with idempotent deduplication and automatic fallback to a manual outbox for platforms without API access.
A Claude Code skill that transforms SRT subtitle files from video monologues or interviews by Wang Jianshuo (็ๅปบ็ก) into multiple standalone WeChat public account articles, one per distinct topic. It identifies independent topics from subtitles, converts spoken language into polished written prose, and creates drafts automatically.
A Claude Code skill for systematically growing X (Twitter) followers through numbered A/B-testable growth experiments, tracking conversion ratios (new followers per profile visit) with built-in rollback and daily CSV-based analytics evaluation.
A skill for systematically improving X (Twitter) content by iterating on content-generation prompts and analyzing per-tweet impression data. Uses git-SHA-versioned prompt experiments with median impressions as the north-star metric to identify which prompt versions and content features produce the highest reach.
A skill for migrating WordPress sites to Hugo static sites on GitHub Pages, taking a WXR export and uploads folder as input to generate Markdown content while preserving URLs, localizing images, and deploying via GitHub Actions.
Skill
Skill
Skill
Skill
Skill
Skill
Skill
Skill
Skill
๐Plugins33
Auto-edit 2+ synced cams into a single MP4. Switches between angles second-by-second based on audio energy. Optional picture-in-picture.
Batch upload videos to YouTube with per-video metadata from `UPLOAD_META.md`. Survives SOCKS/HTTP proxies by using `requests` directly for resumable upload.
Turn a ็ๅปบ็ก-style WeChat article into a narrated 1080ร1920 portrait MP4 (30โ90s) โ Volcano TTS voiceover, HyperFrames CSS/GSAP per-scene animation, subtle SFX, abstract watercolor background.
Audio/video โ timestamped SRT in the source language. Chinese uses Volcano ASR; other languages use OpenAI Whisper with word-level timestamps + punctuation-bounded re-segmentation.
Cut a long-form interview / lecture / podcast into 3โ6 stand-alone topical short clips, with per-clip SRT โ ready for downstream post-production.
Holistic state-of-the-project audit: unmerged branches, stalled PRs, failed GitHub Actions, stale builds, plan drift, unreleased commits, log errors. Read-only investigation โ grouped checklist โ fix only after confirmation.
Add an in-site feedback loop to a website repo โ a floating ๆไธชๅปบ่ฎฎ button where allowlisted visitors' suggestions become GitHub Issues that Claude Code turns into auto-merged, auto-deployed code changes, with a /_feedback dashboard + one-click revert.
Auto-syndicate your newest ๅพฎไฟกๅ ฌไผๅท article across platforms โ auto-posts to X / Bluesky / Threads / LinkedIn and prepares a copy-paste outbox for manual platforms (Facebook / ๅฐ็บขไนฆ / ๅณๅป / ็ฅไน).
Migrate a WordPress site (WXR export + uploads) to a Hugo static site on GitHub Pages โ preserves /archives/<id>/ URLs, localizes images, deploys via GitHub Actions.
Turn a single English word into an HLS supercut lesson from a video base โ stitches every clip where the word is spoken, prepended with a bilingual intro card (word + IPA + ไธญๆ gloss + usage, Volcano TTS) and a follow-me CTA card.
VoiceDrop ้ ๅฅ skillใไธคไธชๆจกๅผ โ (1) list๏ผๅๅบไฝ ๅจ VoiceDrop R2 ไธ็ๅ จ้จๆไปถ๏ผๅฝ้ณ / ๆ็ซ JSON / ๆ ่ฎฐ๏ผ๏ผ(2) distill๏ผไปไฝ ็ๆ ทๆฌๆ็ซ ่ธ้ฆๆ้ฃ๏ผ็ๆๅฏ่ดดๅ ฅ VoiceDropใ่ฎพ็ฝฎโๆ้ฃใ็็ฒพ็ฎ่งๅ้๏ผๅนถๅฏ็ดๆฅ้่ฟ Files API ไธไผ ใ่งฆๅ่ฏ๏ผ"voicedrop list"ใ"ๅๅบ VoiceDrop ๆไปถ"ใ"voicedrop distill"ใ"่ธ้ฆ VoiceDrop ๆ้ฃ"ใ"/wjs-voicedrop"ใ
Delete ๅผๆต / emoji spam replies (ๅๅ้ขไป / ๅฏปๅบ็ฎ / ็บฟไธไธ้จ etc.) under your X posts. Covers the last 7 days via X recent-search.
Turn uploaded VoiceDrop voice memos (the jianshuo.dev/files R2 inbox) into WeChat article drafts โ pull the unprocessed recordings, transcribe, and mine articles from each.
Time-aligned TTS dub of a video against a target-language SRT. Volcano TTS for Chinese, edge-tts for everything else. Single-speaker by default; opt-in visual diarization for multi-speaker.
Post a daily X tweet inspired by your newest un-tweeted WeChat article โ drafts 3 angle variants (quote / metaphor / one-liner), posts the chosen one via xurl, records history.
Mine a video's SRT โ a monologue / ่ฎฒ่งฃ, or an interview where ็ๅปบ็ก is one speaker โ into multiple standalone ๅพฎไฟกๅ ฌไผๅท articles, one per distinct topic.
Conversational backend for a Hugo static blog โ add / edit posts, manage categories, upload images, commit, push, auto-deploy. No CMS, no server.
Ship a TestFlight iOS app to the App Store โ prepare screenshots + description/metadata and submit for review via the existing fastlane setup.
Capture a writer's voice from a few sample articles and rewrite text to sound like that author. Works for any author, ็ๅปบ็ก included.
wjs-evaling-voicedrop-prompts
Audio cross-correlation alignment for N recordings of the same event. Outputs `.sync.json` sidecars per file โ never re-encodes the originals.
Post-production overlays on top of a clip: AI-generated cover as first frame, HTML/CSS kinetic captions, illustration overlays, chapter chips, end-card CTA. HyperFrames-backed, single final encode.
Wire up fastlane + GitHub Actions for an iOS project so pushing main auto-builds and uploads to TestFlight.
Grow X followers via numbered, A/B-testable growth experiments โ each with a hypothesis, target metric, rollback state, and verdict. North-star = new-follower รท profile-visit conversion.
Horizontal โ vertical conversion (16:9 โ 9:16, 4:3 โ 3:4). Crops a narrow band and tracks the active speaker via MediaPipe + mouth-aspect-ratio variance.
Systematically improve X content by A/B-testing the tweet-generation prompt โ git-SHA-versioned, numbered experiments judged on median impressions per tweet, with per-tweet feature mining.
Final-composite step: burn SRT into pixels via libass, mix in a dub track, keep the original audio as a low-volume bed โ all in one ffmpeg encode (no cascade).
Translate an SRT from one language to another with punctuation-bounded re-segmentation. Simplified Chinese and English are first-class targets. Supports bilingual output.
ๅไธๅ ้ฟไธๆบ โ 6-step interactive reflection for personal setbacks. Diagnoses which layer (L1 notes / L2 in-the-moment trigger / L3 default-reaction practice) is missing and routes to the right modality, instead of letting every reflection become another notebook entry.
Thin orchestrator for the end-to-end video localization pipeline (transcribe โ translate โ dub โ burn). Use it when the user asks for full localization in one go.
Rewrite a plain Chinese tweet into a high-engagement version backed by real facts + an interaction hook. Returns 2โ3 variants, short / plain / colloquial.
Polish a ๅพฎไฟกๅ ฌไผๅท (WeChat Official Account) article, generate a cover image (้ขๅพ) and explanation illustration, and upload as a draft to mp.weixin.qq.com.
Automated daily promotion for your Claude Code skills โ picks a skill, generates a marketing angle, auto-posts to X via xurl, and drafts community posts (Reddit / HN / Discord) to an outbox.