agentleFS
harshil-1411harshil-1411/claude_plugin_video_studio/skills/review/SKILL.md

review

Skill0 starsChanged 48 days agoApache-2.0View on GitHub

Look at a rendered video-studio reel as images - a contact sheet of every scene's opening, middle and closing frame, a sheet of every transition at its midpoint, a strip of every frame in a span (motion, transitions, word cues), or full-resolution crops (captions, small text, faces) - and report or fix what is wrong. Use when the user runs /video-studio:review, asks how a render looks, or before handing over a preview or final render.

192 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use close-spec, design-with-images and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • dzhngSkill

    close-spec

    When a spec under specs/<feature>/ is done shipping (a write-spec build, or any planned task that produced a spec), archive it to specs/done/ and rewrite it from a build-plan into a durable rationale — the why, the principles, the invariants — pointing back to the real code for the how. Use when implementation has landed and the plan no longer matches what shipped, or the user says a feature/spec is finished. Pairs with [write-spec](../write-spec/SKILL.md) (the plan this closes) and [review](../review/SKILL.md) (run before closing).

    Also in 5 projects that use this

  • dzhngSkill

    design-with-images

    Explore UI and visual design with image generation, then iterate the real implementation against the user-selected concept. Use when brainstorming visual directions, offering mockup variants, or implementing an approved generated design; do not stop at the mockups or the first code approximation.

    Also in 5 projects that use this

  • LambenthanSkill

    daily-arxiv

    每日从 arXiv 拉取新论文,过滤相关性,auto-ingest 高优先级论文,检测 SOTA 更新

    Also in 5 projects that use this

  • LambenthanSkill

    exp-eval

    实验判决门:Review LLM 独立评判实验结果 → 4 种判决路径 → 自动更新 claims confidence、ideas status、graph edges

    Also in 5 projects that use this

  • +3 more from dzhng
  • +5 more from Lambenthan

The file

We split this skill into the rules it gives an agent

1 rule. 0 are common, 0 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 157 edited versions we read keep these.

  • 0

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Check that the blend reads: no two blocks of text on top of each other, no flash, no half-built layout.

    line 56

“Common” is measured: kept by at least 6 in 10 of the 157 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Review a render

You can't watch the video, but you can look at it. The review tool writes images of labelled frames (scene id, in/mid/out, time) to qa/review/ and returns their paths in images (image is the first). Each image stays within the ~1568 px vision limit so labels stay readable: a long video's contact sheet is split into sheet-<quality>-p1.jpg, -p2.jpg, …, three frames per scene, and pages says which scenes (and which flagged scenes) each image holds. Read every image, flagged ones first.

1. Contact sheet first

mcp__plugin_video-studio_engine__review {project_dir} (the latest render; quality to choose). Review runs lint on the same render first: tiles of scenes with findings get a red (error) or amber (warning) border, and the result's flagged list names each scene's findings (e.g. s03: text_overflow (error)). Look at the flagged tiles first and confirm each finding by eye (the result's pages[].flagged says which image they are on); then check every scene:

  • Text: fits its box, readable at phone size, nothing cut off, no stray characters (e.g. a number glued to its unit).
  • Placement: nothing under the captions or the platform UI.
  • Completeness: by mid the scene's graphic is fully on screen; out shows the settled end state.
  • Footage: crops keep the face or subject; redactions cover what they should.
  • Rhythm: consecutive scenes look different enough (a visual change every 2-4 s), and the hook's first frame already says something.

2. Look closer where something is off

  • mode: "strip", scene: "<id>": every frame of that scene. Use it for motion (push_in, punch, reveal), transitions, count-ups, and word cues: a cued item should appear as its word is spoken. The strip labels the tile nearest each placed cue with cue "<word>" (also in the tile's cues); check the item is appearing on that tile. Unplaced cues show up as cue_unmatched flags. A long span is sampled evenly up to 48 frames; narrow it with from_sec/to_sec for every frame.
  • transitions: true: one tile at the middle of every transition, labelled s02→s03 crossfade @ 12.40s (its transition field gives the kind and length). Cuts get no tile. Tiles use the render's scene slots and the length the assembly actually used: the scene's transition, else the style's default, capped at 40% of the incoming scene and 1.5 s, and a cut when that is under two frames. At the middle both scenes are half visible. Check that the blend reads: no two blocks of text on top of each other, no flash, no half-built layout. A busy blend wants fade_black, a cut, or a stagger (old content out, then new in). Add scene to see only the transitions into and out of one scene, and use a strip over a transition to see every frame. stills can't show transitions: they are drawn at assembly.
  • mode: "crop", crop: {x, y, w, h} (fractions of the frame) with times: full-resolution detail such as the caption band ({x: 0, y: 0.65, w: 1, h: 0.25}), small labels or a face.

3. Report or fix

List what you found per scene, most visible first, with the fix in spec terms (shorter on_screen_text, a different motion, footage.focus, a cue on a later word). When the user asked you to fix things (or you are in the render or create flow), edit the spec, re-render (only changed scenes re-render) and review again. Don't claim a render looks right without having Read every image of the review. To let the user watch two versions (preview and final, or before and after a fix) in sync, suggest the compare skill.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.