More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Stage Assemble: Video Plan to Finished File

    1

    Assemble an approved multi-source video plan into a finished file: checkpointed segments, layered assembly, narration mixed exactly once, and a QA gate before delivery.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Stage Assemble: Video Plan to Finished File

    Stage Assemble: Video Plan to Finished File

    Example session with this skill installed

    Assemble this approved plan: a 60-second vertical launch video, 5 segments (2 footage edits, 1 generated clip, 1 composed title card, 1 provided outro), narration across 4 windows, captions supplied, canvas 1080×1920 at 30 fps. One generated clip is billable.

    • Read your context and instructions
    • Compiled the stage assemble

    RUN

    • Pre-flight: plan validated, caption data final, narration windows do not overlap, all sources present, billable generation count (1) confirmed before anything produces. Journal opened: 5 segments pending.
    • S1 footage edit — probed 14.8 s vs 15.0 planned, on-canvas: verified, warn absorbed by the following card.
    • S2 generated clip — produced once, 6.0 s, silent as required: verified.
    • S3 composed title card — arrived with a voiceover baked into its render: REJECTED at intake under the exactly-once narration rule. Re-rendered silent, journal updated, no other segment touched.
    • S4 footage edit — 12.1 s: verified. S5 provided outro: verified.
    • Assembly, layers in order: primary track conformed to the declared 30 fps canvas, the full-screen card placed in the primary track (not as an overlay), narration mixed exactly once across the 4 windows, one music bed ducked under the voice, captions burned last from caption data, loudness normalized and measured: −14.0 LUFS integrated, −1.0 dBTP true peak.

    QA GATE — Technical: pass (60.2 s, on-canvas, audio present). Promise: pass (real motion 78% of runtime vs 65% committed). Visual: pass, 10 frames pulled across the timeline, no garbled text or empty frames. Audio: pass, narration coverage full, no overlaps, transcript matches the plan.

    DELIVERED — the finished file, the journal (5 verified, 1 rejection, 1 re-render), and this QA report. The one defective segment cost one re-render — not a rebuild of the other four.

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

    About this skill

    The assembly discipline for multi-source video: turning an approved plan — footage edits, generated clips, composed graphics, supplied pieces — into one finished file without losing anything the plan promised. It is built against the three enemies of assembled video: rework, double audio, and silent drift. A checkpoint journal records every segment's status and output path the moment it verifies, so a crash or a revision at 90% costs 10%, not a rebuild; billable generations get an extra authorization guard. Segments are produced by source in plan order and probed on arrival — duration, resolution, audio presence — with numeric intake scoring that rejects off-spec pieces at the door instead of at final QA. Assembly runs in a fixed six-layer order (primary track, overlays, narration exactly once, one continuous music bed, captions burned last from data, loudness normalization with measured numbers) because the order itself is the correctness argument: most assembled-video defects, including the signature two-voices-saying-the-same-line failure, are layer-order violations. A continuity section covers how mixed sources read as one video, a worked run traces a 90-second explainer through one rejection and one send-back, and no draft ships without a written four-section QA report — technical probe, promise check, visual spot-check, audio check — with bounded send-backs to the single segment at fault.

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

    • It assembles an existing approved plan; it does not write the plan or generate footage itself — planning and generation are separate stages.
    • The visual spot-check requires the ability to inspect frames; where that is unavailable the skill marks the check unverified rather than guessing.
    • Loudness targets follow common delivery practice; a specific platform or broadcaster spec you must hit should be stated in the plan.

    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

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 24 days ago

    Permissions

    No special permissions declared or detected

    Listed24 days ago
    Updated21 days ago

    What's inside

    Frequently Asked Questions

    Because the classic assembled-video defects are order violations: narration baked into a segment and mixed again produces two voices on one line, and captions burned early turn a typo fix into a full re-render. The six-layer order makes both defects structurally impossible instead of relying on vigilance.

    A bounded send-back to that segment only. The checkpoint journal proves which segments are verified, so they are never re-produced; allow at most two send-back rounds on the same check before the options come back to you with the artifact preserved.

    No. It sits above them: specialist steps still produce each segment, and your tooling still renders. This skill receives and verifies their outputs, enforces assembly order, and gates delivery — the coordination layer where multi-source projects usually fall apart.

    New in Image & Video