
video-publish
Gets a finished video out on YouTube (Shorts or long), Instagram Reels and TikTok - picks the video, makes the light export for Instagram and TikTok, the cover, the caption and the YouTube title, sets the time, shows the plan, then after the creator's yes publishes through Treg on their own connected accounts (setup included), or hands a ready-to-post kit when Treg isn't set up. Use when the user says publish, post it, schedule, upload to YouTube, put it on Instagram, TikTok, publier, poster, programmer, mets-le en ligne.
- Installs
- 46
- Last 24h
- —
- Trending rank
- —
- Security
- Not audited
Install
Works with every agent the skills CLI supports.
npx skills add https://github.com/melvynx/video-agent-skills --skill video-publish- 01
Install
Run the command in your project. The CLI asks which agents to install for.
- 02
Describe the task
Ask your agent for something this skill covers. It loads video-publish automatically when relevant.
- 03
Review the result
The agent follows the SKILL.md steps, scripts and references bundled with the skill.
SKILL.md
The instructions your agent reads when the skill is triggered.
Video Publish
One finished video out, on the creator's own accounts. Making it is /video-edit.
Nothing is posted before the creator's explicit yes on the plan of step 5. That yes covers that video only.
Talk to them in their language. <edit> = video-agent/edits/<slug>/.
1. Which video
First match wins:
- A path, a slug or a title they give:
<edit>/final.mp4. - The video finished in this conversation.
- Nothing: the newest
video-agent/edits/*/final.mp4. Name it in the plan.
Then check <edit>/published.md: already there with this platform, say so and stop (unless they want to post it again).
Read the format with ffprobe: vertical (a short, for YouTube Shorts, Reels and TikTok) or horizontal (a long video, YouTube only by default).
2. Setup Treg (once)
Treg posts for them on their own accounts, free. Check:
treg --version
treg connections ls
Each platform they post on must show in the list with posting rights (scopes with content_publish on Instagram, video.publish on TikTok, youtube.upload or youtube on YouTube).
Missing: explain in one line, ask, then:
- Install (macOS and Linux; on Windows use the MCP:
claude mcp add --transport http treg https://treg.to/mcp/, then/mcpto sign in):curl -fsSL https://treg.to/install.sh | sh treg login - Connect only the platforms they post on, each opens the platform's own login:
Instagram needs a professional account (Creator or Business, free switch in the app's settings).
treg connections connect --provider instagram --capability post treg connections connect --provider tiktok --capability post treg connections connect --provider youtube --capability post
They don't want Treg: the manual route (6a) works with no setup.
3. The files
- Master:
<edit>/final.mp4, full quality, for YouTube. - Social export (shorts only):
<edit>/social.mp4, H.264 + AAC, under 30 MB so Treg can host it for Instagram (TikTok accepts up to 64 MB). Missing or older than the master, re-encode at a bitrate that fits (about 28 MB × 8000 ÷ duration in seconds, minus 160 kb/s of audio, in kb/s):Check its size, and that it decodes:ffmpeg -y -i <edit>/final.mp4 -c:v libx264 -preset slow -b:v <rate>k -maxrate <rate>k -bufsize <2x rate>k -pix_fmt yuv420p -c:a aac -b:a 160k -movflags +faststart <edit>/social.mp4ffmpeg -v error -i <edit>/social.mp4 -f null -. - Cover:
<edit>/cover.jpg, always. Without one, Instagram takes frame 0, often a blink or a blank frame. Pick it from the hook: the face with the eyes open, the hook card fully in. Contact sheet of the first 3 seconds, read it, then extract the chosen time<t>:Vertical: keep the key visual inside the central 3:4 (y 240 to 1680), the profile grid crops to it. A long video: a 1280x720 thumbnail is better; offerffmpeg -v error -y -ss 0.1 -t 3 -i <edit>/final.mp4 -vf "fps=10,scale=270:-1,tile=6x5" -frames:v 1 video-agent/.tmp/hook-sheet.jpg ffmpeg -v error -y -ss <t> -i <edit>/final.mp4 -frames:v 1 -q:v 2 <edit>/cover.jpg/codex-imagesto make one, or use a frame.
4. The texts
From <edit>/post.md (written by video-edit) when it exists, else from the video's script or transcript and video-agent/voice.md, in the video's language:
- Caption (Instagram and TikTok): a first line that hooks without repeating the spoken hook word for word (the first 125 characters show before "more"), 2 to 4 short lines on what the viewer gets, the ask from voice.md, then 3 to 5 hashtags of the topic. No #fyp, no wall of emoji.
- TikTok caption only when it should differ (shorter, other hashtags).
- YouTube title: Shorts under 60 characters, long under 70, the promise of the video, true to it.
- YouTube description: the caption, the chapters for a long video (from post.md), then their links (ask once and add them to voice.md).
Run the humanize pass of video-script (its references/humanize.md) on all of them when that skill is installed. Every name, number and claim comes from the video; none is added. Save them in <edit>/post.md, one heading per text.
5. The plan (mandatory)
- Platforms: a short goes to YouTube + Instagram + TikTok, a long video to YouTube, unless they name others.
- Time: the one they give (written in their local time and in UTC), else now.
- Route per platform: Treg (connected) or manual. Instagram and TikTok have no scheduling through Treg: they post now. YouTube can wait for a date (private +
publishAt). A later time on Instagram or TikTok: run/video-publishagain at that time, or go manual and use the app's scheduler.
Show it and wait for a clear yes:
PLAN · budget-margins · 0:47 · social.mp4 18 MB
Cover: cover.jpg (0.8 s) [shown as an image]
YouTube Shorts · Treg · 2026-10-11 18:00 Paris (16:00 UTC) · private until then
Instagram Reel · Treg · now
TikTok · manual (not connected)
Title: Your margin is lying to you
Caption: ...
Reply yes to publish, or tell me what to change.
A change on the plan: apply it, show the plan again.
6a. Manual
Open the folder (macOS open <edit>, Windows explorer <edit>, Linux xdg-open <edit>), then give the steps:
- YouTube: studio.youtube.com → Create → Upload videos →
final.mp4→ title and description frompost.md,cover.jpgas thumbnail → Visibility → Schedule or Public. - Instagram: the app → + → Reel →
social.mp4→ the caption → Edit cover → Add from camera roll →cover.jpg→ Share (or Schedule in the advanced settings of a professional account). - TikTok: the app or tiktok.com/upload →
social.mp4→ the caption → cover at the cover time → Post (or Schedule on the web).
6b. Treg
Before each call, treg catalog get <id> for the exact parameters (they change). All of these are free on their own accounts. Keep stdout only (2>/dev/null). One platform failing never blocks the others.
Instagram reel
- Host the files (30 MB max each, 7 days):
treg host <edit>/social.mp4andtreg host <edit>/cover.jpgprint public URLs. instagram.instagram.media.container.createwithig_user_id(theresource_refof their Instagram intreg connections ls),media_type=REELS,video_url,cover_url,caption.- Poll
instagram.instagram.media.container.statusevery 10 s untilFINISHED(ERROR: show it and stop for Instagram). instagram.instagram.post.publishwith the container id. Then itspermalink(instagram.instagram.user.postswithfields=id,permalink, the newest one).
TikTok
tiktok.tiktok.creator.info: the privacy options they may use.tiktok.tiktok.video.publish.initwith the caption, aprivacy_levelfrom that list (public when it is there),video_cover_timestamp_ms= the cover time in ms (TikTok takes a frame, never an image), andsource_infoFILE_UPLOADwith the file size as one chunk.- PUT
social.mp4to the returnedupload_urlwithContent-Type: video/mp4andContent-Range: bytes 0-<size-1>/<size>. - Poll
tiktok.tiktok.publish.statusuntil it is published.
An answer unaudited_client_can_only_post_to_private_accounts (or only SELF_ONLY in the options): send it to their TikTok inbox instead with tiktok.tiktok.video.inbox.init (same upload, no caption), and tell them to open TikTok within 24 hours, paste the caption and post.
YouTube
youtube.youtube.video.uploadwithuploadType=resumable, the snippet (title, description) andstatus{"privacyStatus": "private", "selfDeclaredMadeForKids": false}plus"publishAt": "<UTC ISO time>"for a date ("public"for now).- PUT
final.mp4to the session URL it returns. youtube.youtube.video.thumbnail.setwithcover.jpg(works on verified YouTube accounts; a refusal is fine, say it).
No session URL in the answer: go manual for YouTube.
7. Record it
Write <edit>/published.md, one line per platform: date, time, route, and the link (https://youtube.com/shorts/<id> or https://youtu.be/<id>, the reel, the TikTok) once live. Manual route: write it when they say it is posted.
Then say what is live, the links, and offer /video-audit in a week to see how it did.
Honest limits
- Instagram and TikTok can't be scheduled through Treg, only posted now. Their own apps can schedule.
- TikTok keeps posts from new third-party apps private until the app is audited; the inbox route works around it.
- A platform can refuse a video (length, format, music rights). The error is shown as is; the manual route always works.
- Only post videos the creator owns or has the rights to, music included.
FAQ
What does the video-publish skill do?
Gets a finished video out on YouTube (Shorts or long), Instagram Reels and TikTok - picks the video, makes the light export for Instagram and TikTok, the cover, the caption and the YouTube title, sets the time, shows the plan, then after the creator's yes publishes through Treg on their own connected accounts (setup included), or hands a ready-to-post kit when Treg isn't set up. Use when the user says publish, post it, schedule, upload to YouTube, put it on Instagram, TikTok, publier, poster, programmer, mets-le en ligne.
How do I install video-publish?
Run npx skills add https://github.com/melvynx/video-agent-skills --skill video-publish in your project. The skills CLI detects the agents you use and copies the skill into the right folder. For Claude Code you can also place the skill folder at ~/.claude/skills/video-publish/ (personal) or .claude/skills/video-publish/ (project).
Which AI agents can use video-publish?
Any agent that supports the open Agent Skills format (a folder with a SKILL.md file), including Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, Windsurf, Cline, OpenCode and more.
Is video-publish safe to install?
No partner security audit is available yet. Skills can run commands and scripts, so read the SKILL.md and any bundled files before installing.
Who maintains video-publish?
It is published by melvynx in melvynx/video-agent-skills. It has 46 recorded installs.
More from melvynx/video-agent-skills
short-editing
melvynx/video-agent-skills
Edits a talking-head take end to end into a paper-style motion design short (1080x1920, also 1:1 and 16:9) with its own render engine (Node + Chrome + ffmpeg, free, offline). Use when the user says edit, cut, captions, subtitles, motion design, montage, "make it a short", render, or gives a recording from video-agent/takes/. Transcribes locally with word timings (whisper.cpp, or ElevenLabs when a key is set), cuts every silence, lays the boards on a textured paper above the camera card (counters, terminals, logos, trees, real screens, HTML motion components), word-by-word captions, a calm music bed, a few designed hits, -14 LUFS, and writes the mp4, an .srt and the post copy.
codex-images
melvynx/video-agent-skills
Open the skill to read its SKILL.md instructions.
studio-script
melvynx/video-agent-skills
Open the skill to read its SKILL.md instructions.
studio-setup
melvynx/video-agent-skills
Open the skill to read its SKILL.md instructions.
steal
melvynx/video-agent-skills
Open the skill to read its SKILL.md instructions.
clips
melvynx/video-agent-skills
Turns one long video (YouTube link, podcast, webinar, livestream, or a local file) into finished vertical shorts. Use when the user says long to short, repurpose, long video to shorts, clips from my video, podcast clips, cut my YouTube video into shorts, or gives a long video link. Finds clips that stand alone (one idea, 20 to 60 seconds), scores them, cuts them and edits each one with short-editing (split-screen paper look, captions, cuts, hook, boards, 1080x1920), and writes new scripts in the studio-script format from the ideas too spread out to clip. Reads YouTube captions and chapters with yt-dlp, or transcribes a local file with the short-editing engine (whisper.cpp, free, offline).
Popular media & creative skills
hyperframes
heygen-com/hyperframes
Read first to make, edit, animate or render any video with HyperFrames (HTML rendered to video), including launch films, explainers, music videos, charts, maps, vertical clips, title cards, captions and overlays. Holds the render contract, the work loop and what each kind of video needs.
media-use
heygen-com/hyperframes
Get or make any media for a HyperFrames video - music, sound effects, images, icons, logos, voiceover, avatar and generated video, transcripts, cutouts, colour grades and looks - with HeyGen or free local tools, and mix placed audio. Also for footage that looks dark, flat or dull, or should be stylized or hidden.
hyperframes-registry
heygen-com/hyperframes
About 450 ready-made HyperFrames blocks and components installed as editable source - transitions, overlays, titles, kinetic type, captions, charts, map overlays, 3D, camera moves. Use it before hand-building any named visual (CRT scanlines, glitch, film grain, a chart); also for hyperframes add and catalog.
hyperframes-creative
heygen-com/hyperframes
Non-animation creative direction for HyperFrames videos. Use for design spec (frame.md / design.md) handling, palettes, typography, narration, beat planning, audio-reactive visuals, composition patterns, and brand / style decisions. For atomic motion patterns and scene blueprints, use hyperframes-animation.
ai-video-generation
101-skills/superpowers
Generate AI videos with Google Veo, Seedance 2.0, HappyHorse, Wan, Grok and 40+ models via inference.sh CLI. Models: Veo 3.1, Veo 3, Seedance 2.0, HappyHorse 1.0, Wan 2.5, Grok Imagine Video, OmniHuman, Fabric, HunyuanVideo. Capabilities: text-to-video, image-to-video, reference-to-video, video editing, lipsync, avatar animation, video upscaling, foley sound. Use for: social media videos, marketing content, explainer videos, product demos, AI avatars. Triggers: video generation, ai video, text to video, image to video, veo, animate image, video from image, ai animation, video generator, generate video, t2v, i2v, ai video maker, create video with ai, runway alternative, pika alternative, sora alternative, kling alternative, seedance, happyhorse
ai-image-generation
101-skills/superpowers
Generate AI images with GPT-Image-2, FLUX, Gemini, Grok, Seedream, Reve and 50+ models via inference.sh CLI. Models: GPT-Image-2, FLUX Dev LoRA, FLUX.2 Klein LoRA, Gemini 3 Pro Image, Grok Imagine, Seedream 4.5, Reve, ImagineArt. Capabilities: text-to-image, image-to-image, inpainting, LoRA, image editing, upscaling, text rendering. Use for: AI art, product mockups, concept art, social media graphics, marketing visuals, illustrations. Triggers: flux, image generation, ai image, text to image, stable diffusion, generate image, ai art, midjourney alternative, dall-e alternative, text2img, t2i, image generator, ai picture, create image with ai, generative ai, ai illustration, grok image, gemini image, gpt image, openai image, chatgpt image