More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Short-video pipeline: topic → script → finished cut

    5

    Four skills that take a news beat from "what should I make today" to a rendered, captioned, publishable video — with two deliberate human checkpoints.

    $15

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    7 uses5.0 (1 review)
    Short-video pipeline: topic → script → finished cut

    Short-video pipeline: topic → script → finished cut

    Example session with this skill installed

    SpaceX just flew its first operational batch of flat-packed Starlink V3 satellites and Musk put a $200B revenue number on it. Make me a video — long cut and a short cut.

    • Read your context and instructions
    • Compiled the short-video pipeline

    STAGE 1 — TITLE (checkpoint, waits for you)
    Three options, each committing the piece to a different angle:
    A. "The satellites are real. The $200 billion is arithmetic."
    B. "Musk's newest number, checked line by line"
    C. "What flat-packing a satellite actually buys you"
    → you pick. Everything downstream follows the angle you chose.

    STAGE 2 — SCRIPT
    733 words, three-part structure. Facts cross-checked against the primary
    source before they reach the page. On this story that meant:
    · two outlets gave conflicting cumulative sales figures → NEITHER went in
    · a projected number was labelled "planned capacity" every time, never output
    · the transcript of the talk was pulled and read, not an aggregator's summary
    of it — which is where the most useful line turned out to be

    STAGE 3 — SIGN-OFF (checkpoint, waits for you)

    STAGE 4 — VOICEOVER
    edge-tts → assets/narration.mp3 (3:52), captions.vtt, word-level cues.
    Completeness check: captions vs script. It has caught a real truncation —
    92s of narration where 161s were expected, exit code 0, no error anywhere.

    STAGE 5 — FOOTAGE
    First-party sources preferred, then every clip inspected across the span it
    will actually occupy, with the crop that will actually be applied:
    · one clip's title promised deployment footage; sampling its full 11 minutes
    showed it never shows the deployment → dropped
    · the best bay footage was from an earlier flight carrying SIMULATORS →
    used, but labelled on screen as exactly that, and only under narration
    about the mechanism, never under the sentence about the real launch

    STAGE 6 — SCENE PLAN → plan.json, plan-short.json
    Boundaries taken from caption cue edges, never rounded.

    STAGE 7 — RENDER (silent → audio → final, so a caption fix is not a re-encode)
    renders/starlink-v3-16x9-final.mp4 3:52 1920x1080 34 scenes
    renders/starlink-v3-short-9x16-final.mp4 2:41 1080x1920 26 scenes

    STAGE 8 — QA
    Stills extracted and read. On this video they caught an on-screen label
    reading "Gen 2" while the narration discussed the third generation.
    ffprobe reports that file as perfectly healthy.

    STAGE 9 — COVER + DISTRIBUTION COPY (+ optional YouTube upload)

    Both finished cuts from this run are public:
    https://www.youtube.com/watch?v=RIgpSjUw_hM
    https://www.youtube.com/watch?v=sPMsLv1DURI

    plan.json, plan-short.json

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Most video tooling starts after you already know what to say. This starts before that, and stops at a finished MP4.

    topic-scan — scans your source list in priority order, verifies every URL resolves, scores candidates on reach, engagement, freshness and fit, and returns a ranked Top 6 with the lead flagged.

    script-writing — five-step pipeline: title, frame, research, draft, sign-off. Cross-checks facts against at least three independent outlets before they reach the page. Picks one of five category templates (opinion, event, product, showcase, incident) and applies that category's cold open, analogy system and closing form. Produces a ~750-word three-part voiceover.

    news-short-video — TTS, sourced b-roll, data and quote cards, burned-in subtitles, QA stills, covers, distribution copy, optional YouTube upload. Nine scripts, three render tiers, 16:9 and 9:16.

    video-pipeline — chains the three and owns the seams between them. The seams are where these things actually break: facts arriving unverified, a word count line getting read aloud by the TTS, a script written to the wrong length for the platform it is going to.

    What you configure

    Three files, once. Persona and beat, your source list, 20–50 filter keywords. The method is fixed; the beat is yours. Switching from EV to biotech means editing those three files and nothing else.

    The two checkpoints are deliberate

    You pick the title, and you sign off the script. Everything in between — the research, the scene plan, the render, the QA stills, the cover — the agent does. Those two stay with you because the title sets the angle of the whole piece and the script carries the factual liability. This is not one-click video, and it is not trying to be.

    What it is honest about

    It orchestrates external tools it does not bundle: ffmpeg, yt-dlp, edge-tts, Node. Run the included doctor.mjs first — one command tells you exactly what is missing and how to install it on your platform. Several of these dependencies fail silently (a missing font makes ffmpeg substitute a face and return success), which is why the preflight exists and why it exits non-zero.

    Footage rights are yours. It fetches from public channels for commentary and citation and prefers first-party sources. Clearing that for commercial publication is your call.

    The voiceover is synthetic. Every platform has a disclosure control.

    Ships configured for English, but not limited to it. Another language needs a font that has the glyphs and the matching voice — both settings. For a script without spaces, keep caption cues inside the line limit; the line-breaker splits on whitespace. TROUBLESHOOTING.md covers both.

    Track record

    This runs a live channel, not a demo account. One cut produced with it has passed 130,000 views on RedNote. That one was Chinese-language; the two worked examples below are the English build, both public, both finished in a single pass through the pipeline. Watch them before you buy — they are the honest output, not a showreel.

    Long cut, 16:9, 3:52 — https://www.youtube.com/watch?v=RIgpSjUw_hM Short cut, 9:16, 2:41 — https://www.youtube.com/watch?v=sPMsLv1DURI

    What ships

    33 files. Four SKILL.md definitions, nine Node scripts, reference docs for the scene-plan format, dependencies, upload and distribution copy, a TROUBLESHOOTING.md for the failure modes that return exit code 0 and pass ffprobe cleanly, and a commercial licence.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Video Skills.

    • Ships configured for English, not limited to it. Another language needs a font that has the glyphs and the matching TTS voice — both are settings, not code changes. For a script without spaces (Chinese, Japanese) the caption line-breaker will not wrap for you, so cues have to be authored inside the line limit: 30 characters at 9:16, 40 at 16:9. • It does not bundle ffmpeg, ffprobe, yt-dlp, edge-tts or Node; it orchestrates them. Several fail silently, which is why doctor.mjs exists and exits non-zero. • Footage rights are yours. It prefers first-party sources and credits them on screen, but clearing a clip for commercial publication is your call. • The voiceover is synthetic. Every platform has a disclosure control and you should use it. • Not one-click, by design. You pick the title and you sign off the script — two stops that carry the angle and the factual liability. • Developed and run on macOS. Font paths are configurable, but the defaults are macOS system faces; on Linux set HF_FONT_DIR or edit the font table in styles.mjs.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    5.0
    1 review
    5
    1
    4
    0
    3
    0
    2
    0
    1
    0

    This skill has been used 7 times.

    Trust & safety

    Security scan

    Security scan passed 3 months ago

    Software
    Node.jsedge-ttsuv (Python tool runner)hyperframescannot

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Browser
    Network Access
    Environment Variables

    File Scopes

    assets/**
    renders/**
    plan*.json
    script*.txt
    captions*.vtt

    Terminal, network and file access are the core of what this does: it shells out to ffmpeg, ffprobe, yt-dlp and edge-tts, downloads sourced footage, and writes narration, scene plans, renders and QA stills into the video project directory. Browser access is used by the topic-scan stage to open pages that web_fetch cannot read. The host list is not closed, and it would be dishonest to present it as one. topic-scan and script-writing fetch news pages from whatever source list YOU configure, so the outlets reached are yours, not ours. The hosts listed above are the ones the package itself names. edge-tts additionally reaches a Microsoft speech endpoint; it runs as a separate binary via uvx, and this package does not hardcode that address. Environment variables: two optional overrides only, HF_FONT_DIR and HF_VERSION. No credentials are read from the environment, and none are bundled. If you enable the optional YouTube upload you supply your own OAuth client, and it stays on your machine. Note on auto-detection: the scanner does not read .mjs files, so it under-reports this package's nine Node scripts. The declarations above were written by hand against the actual source.

    Listed3 months ago

    What's inside

    Frequently Asked Questions

    No, and deliberately not. You pick the title and you sign off the script. Everything between and after those two points — research, voiceover, footage, scene plan, render, QA stills, cover, distribution copy — the agent does. Those two stay with you because the title sets the angle of the whole piece and the script carries the factual liability. If you want a tool that publishes without you reading it, this is the wrong one.

    Yes. Nothing about the beat is hardcoded. Three config files, edited once: persona and positioning, your source list, and 20–50 filter keywords. The method is fixed; the beat is yours. Switching from EV coverage to biotech means editing those three files and nothing else.

    Yes, with configuration. We run a Chinese channel on this pipeline — one of those cuts passed 130,000 views. Two things to set: point the plan's font at a face that has the glyphs, because ffmpeg's drawtext does not fall back when handed an explicit fontfile and will draw empty boxes while still exiting 0; and choose the matching voice, for example zh-CN-YunxiNeural. One thing to work around: the caption line-breaker splits on spaces, so in a script that has none, keep each cue inside the line limit (30 characters at 9:16, 40 at 16:9) instead of relying on it to wrap — a cue under that limit passes through untouched. What ships is set up for English: the docs, the config templates and the default voice.

    Node 18+, ffmpeg (with libfreetype and libx264), ffprobe, yt-dlp and edge-tts. All free, all open source, no API keys and no accounts. Run the included doctor.mjs before anything else — one command tells you exactly what is missing and how to install it, and it exits non-zero so a broken setup cannot slip past. It checks fonts too, because a missing font makes ffmpeg substitute a face and return success.

    Footage rights are yours to clear. The pipeline fetches from public channels for commentary and citation, prefers first-party sources, and credits the source on screen — but whether a given clip is cleared for your commercial use is your call, not something a script can decide. The voiceover is edge-tts, which is free and runs on your machine; it is synthetic, every platform has a disclosure control, and you should use it. What you are buying is the method and the code, under a commercial licence included in the package.

    Because they activate at different moments and you will often want just one. topic-scan when you are deciding what to cover, script-writing when you already have the story, news-short-video when you already have the script. The fourth, video-pipeline, chains them and owns the seams — which is where this kind of thing actually breaks: facts arriving unverified, a word-count line getting read aloud by the TTS, a script written to the wrong length for the platform it is going to. All four install together from one archive into ~/.claude/skills.

    New in Image & Video