agentleFS
murathany7murathany7/reel-studio/.agents/skills/review-desk/SKILL.md

review-desk

Skill5 starsChanged 8 days agoMITView on GitHub

Optional. Open the Reel Studio desk (a local review app) and work from its review notes - notes pinned to frames, renders, approvals. Use only when the person asks for the desk or already uses it - "open the desk", "apply my desk notes", "fix what I marked", "what's on this frame" - or a production has a review.json. Never required: reviewing and approving in chat works the same.

  • Reads credentials

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 1

    1. 5hseok/hanipman-gate★ 0same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository9 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    When a storyboard is ready, say "it's in the desk for approval" and stop (golden rule 2).

    line 65

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “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 desk (optional)

The kit works the same without it: storyboards are approved in chat, and QA is the make-video steps. Start the desk when the person asks, or they already use it (a review.json in the production). make-video offers it once, when a storyboard is ready; don't push it beyond that.

The desk (desk/) is a local page for the person: it plays a production's composition frame-exact through the renderer's own clock (__seek), and lets them pin notes to frames, render stills, preview videos and QA, read the storyboard, approve, and send notes to a coding agent. It only reads and writes files in the project, so everything it does reaches you as files (and, if it is registered, through the reel-desk MCP server).

Open it

Open it for the person; don't hand them a command to run (many are not technical).

  • Claude desktop app (a preview_start tool is available): preview_start with name desk. It starts the server from .claude/launch.json and shows the desk in the app's browser pane, next to the chat. setup writes that entry; if it is missing, add it: {"name": "desk", "runtimeExecutable": "node", "runtimeArgs": ["desk/server.mjs", "--port", "4100"], "port": 4100} under configurations.
  • Elsewhere (terminal, Codex, Copilot): run node desk/server.mjs in the background and give the person http://127.0.0.1:4100. npx reel-studio desk does the same. Port: --port <n>.

Then open the production they are working on, and say in one line what they can do there: play it, pin notes to a frame, approve the storyboard.

The reel-desk MCP tools below are not registered by default. If they aren't available and the person wants them: Claude Code: claude mcp add reel-desk --scope project -- node desk/mcp.mjs; Codex: codex mcp add reel-desk -- node desk/mcp.mjs; Copilot: add it to .vscode/mcp.json or the Copilot CLI's /mcp add (then restart the session). Without them, read and write review.json directly.

Where things are

Notes and approvalsproductions/<brand>/<ID>/review.json (tracked in git: it is the review record)
What the person has opendesk/.state.json, or the desk_state tool
Desk rendersproductions/<brand>/<ID>/qa/desk/ (stills, preview.mp4, qa/); never the delivered export

Working from notes

  1. desk_state tells you which production and time "this frame" means. list_notes gives the open notes (id, t in seconds, text). Without the MCP server, read review.json.
  2. For each note: change the composition, check it with render_frame at the note's t (or node core/render.mjs still … --t <t>) and look at the image.
  3. resolve_note with a one-line reply saying what changed. If a note needs a fact, an asset or a decision you don't have, keep it open and put the question in the reply. Only the person closes notes (after checking).
  4. Your own QA findings can go on the timeline too: add_note (they show as by claude, the desk's name for any agent).

The desk's Send to agent button runs exactly this loop headlessly with the person's own login: Claude Code (claude -p, ANTHROPIC_API_KEY removed, tools limited to read/edit/render/QA and reel-desk) or Codex (codex exec, workspace-write sandbox, OPENAI_API_KEY removed), picked with REEL_AGENT. One session per production (review.json → claude.session / codex.session).

Approvals are the person's

The desk's Approve storyboard ticks - [x] Storyboard approved in the production README and fills Approved by / date; Approve final cut adds a final-cut line. Both are logged in review.json → approvals. Never write an approval yourself, in the README or in review.json, and never treat a fixed note as approval. When a storyboard is ready, say "it's in the desk for approval" and stop (golden rule 2).

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.