Communitygithub.com

edit-delivery-qa

Verify final video edits before delivery. Use when an agent needs to inspect a rendered MP4/MOV/WebM, confirm duration/resolution/frame rate/audio, generate or review contact sheets, analyze timeline CSVs, check cut density/source mix, check FX integration and cohesion, verify text-to-beat sync, catch weak frames or accidental placeholders, and decide whether a beat-synced social edit is ready to show.

What is edit-delivery-qa?

edit-delivery-qa is a Claude Code agent skill that verify final video edits before delivery. Use when an agent needs to inspect a rendered MP4/MOV/WebM, confirm duration/resolution/frame rate/audio, generate or review contact sheets, analyze timeline CSVs, check cut density/source mix, check FX integration and cohesion, verify text-to-beat sync, catch weak frames or accidental placeholders, and decide whether a beat-synced social edit is ready to show.

Works with~Claude Code~Codex CLI~Cursor
npx skills add https://github.com/BusyBee3333/viral-video-edit-skills/tree/main/skills/edit-delivery-qa

Ask in your favorite AI

Open a new chat with this agent skill pre-loaded.

Documentation

Edit Delivery QA

Overview

Use this skill before presenting a video edit as finished. The expected output is a concise QA readout: technical specs, timeline health, visual review notes, and any fixes needed before delivery.

Required Checks

  1. Probe the final render:
    • duration
    • resolution
    • frame rate
    • video codec
    • audio stream presence
    • delivery container
  2. Generate a contact sheet if one does not already exist.
  3. Review the contact sheet for:
    • active, intentional frames
    • no title cards/interviews/blank frames unless deliberate
    • no watermarks or stock-preview overlays
    • readable subject/action after crop
    • accents that look intentional, not random filler
  4. Inspect the timeline CSV if available:
    • segment count
    • average/min/max segment duration
    • source mix
    • kind/category mix
    • priority-hit usage
  5. Compare the result to the target format and creative brief.
  6. State whether the edit is ready, needs a small revision, or needs a rerender.

Beat-Synced Montage Targets

For dense social montage cuts:

  • Duration: usually 30-40s.
  • Vertical: 1080x1920; landscape: 1920x1080.
  • Frame rate: 30fps minimum.
  • For a 35s high-energy montage, 70-90 segments is healthy if the footage can support it.
  • Average segment length: about 0.35-0.50s.
  • Audio should be present and clean.
  • Contact sheet should show mostly active/intended footage.

Integrated-FX & Craft Checks (2026-07 additions)

For edits with effects or text layers, add these to the required checks:

  • Cohesion / sticker check: does any composited element read as an untracked sticker rather than an object in the scene? Does any segment read as a cutaway to a different video? Does one shot break the shared grade (a one-off look among naturals)? Any yes = revision — these are the highest-frequency rejection causes.
  • Text-sync check: if a hook line is meant to land on a beat, measure the actual offset between the text-appearance frame and the audio accent; |offset| > 100ms = re-render.
  • Loop check: replay the ending into the start; loop-seam edits must be invisible.
  • Phone-scale check: inspect at full size AND phone scale; save review frames at mandated timestamps (e.g. 1/4/8/12/14s of a 15s piece) plus a late-frame spot check alongside the QA report.
  • Cheap-proof stage (before full builds, not after): render a low-res skeleton of the first 3-5 seconds and watch it at scroll speed before committing to the full/VFX render — kill criteria are cheap at this gate and expensive after VFX.
  • QuickTime deliverable: when the review happens in Apple players, also produce and separately QA a QuickTime-compatible .mov.
  • For anything going public, run five independent gates: technical, visual, narrative, originality/rights, platform package. Automated QA supports but never replaces watching the render at normal speed. Grain is fine (viewers prefer it); blockiness, banding, upscale blur, watermarks, and borders are the real penalties.

Failure Patterns

Send the edit back for revision when you see:

  • no audio stream or wrong audio excerpt
  • wrong aspect ratio or unintended crop
  • empty frames, title cards, interviews, stock watermarks, or price overlays
  • action hidden by the vertical crop
  • too many weak crowd/laser/nature inserts replacing the main premise
  • priority musical hits served by low-motion footage
  • timeline CSV and final render disagree materially

Script

Run scripts/qa_video_edit.py against the final video. Add --timeline and --contact-sheet when available:

python3 scripts/qa_video_edit.py final.mp4 --timeline final_timeline.csv --contact-sheet final_contact_sheet.jpg --out final_qa

The script writes JSON and Markdown. It can also create a contact sheet with FFmpeg when --make-contact-sheet is set.

Human Review

The script catches specs and timeline problems; it does not replace watching or inspecting frames. Always pair automated QA with at least a contact-sheet review, and watch the final render when the edit is high stakes or visually dense.

For the Conteto Machino standard, read references/edit-qa-standard.md.

Optional references:

  • references/qa-thresholds.md: adaptable thresholds by edit type.
  • references/example-prompts.md: reusable final-QA prompts.
  • references/gui-vision.md: product vision for a visual QA and delivery panel.

Individual skills in this repo

This repo contains 4 individual skills — each has its own dedicated page.

beat-synced-montage-builder

Build beat-synced social video montages from song-analysis outputs and source footage. Use when an agent needs to assemble, plan, or revise a music-timed edit; map footage to beats/onsets/cut markers; create vertical or landscape montage timelines; integrate audio-reactive FX into shots; mix gameplay, crowd, concert, nature, menu, or other accent footage; generate timeline CSVs, contact sheets, and edit briefs; or adapt the Conteto Machino / Smash Mash workflow for another song or footage library.

generated-media-sourcing

Generate or plan media assets for beat-synced edits when real footage is missing or the user asks for generated visuals - AI video (Gemini/Veo/Flow), AI images, or procedural audio-reactive GPU renders (tunnels, portals, kaleidoscopes, starbursts, mandalas, title pieces via Python+moderngl GLSL). Use with beat-synced-montage-builder when a montage needs generated clips, stills, or procedural audio-reactive elements instead of, or alongside, sourced footage.

song-edit-analysis

Analyze songs and audio files for beat-synced video editing, clip timing, energy mapping, viral window selection, onset-locked cut grids, footage selection, and agent-readable edit briefs. Use when an agent needs to inspect an MP3/WAV/M4A/audio track, estimate tempo and beat grids, identify sections, produce cut markers, pick the best 15-30s window of a song for short-form video, or guide a human or another agent in choosing footage that matches musical energy.

viral-video-edits

End-to-end system for making S-tier beat-synced short-form music video edits — song analysis, viral window selection, footage sourcing with shot-level QA, hook writing, onset-locked cutting, integrated audio-reactive FX (person mattes, procedural GLSL renderers, cinematic grade), and delivery QA. Use this whenever the user asks to make a video edit, music video, reel, TikTok/Shorts/vertical video, beat-synced montage, lyric/hook video, audio-reactive visual, tunnel/portal/kaleidoscope render, or to promote a song with video content — even if they don

Related Skills