working-practice

Guides AI assistants on collaboration rules for building videos with Remotion.

Updated Sep 22, 2026
One-click install
npx skills add https://github.com/napatne/LIVE-Workspace --skill working-practice-napatne
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: working-practice
Source: https://github.com/napatne/LIVE-Workspace/tree/main/.agents/skills/working-practice
Command: npx skills add https://github.com/napatne/LIVE-Workspace --skill working-practice-napatne

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? When an AI assistant builds a video in code, it routinely oversteps: judging creative quality it cannot judge, treating estimated durations as facts, inventing constraints, and writing notes that make no sense to the next reader. This Skill defines the working rules that prevent those failures across every stage of a video project. ## Core Features & Use Cases - Decision Boundaries: Separates what the assistant may verify (rendering, resolution, safe areas, file agreement) from what the human judges (pacing, humour, composition, meaning). - Timing Discipline: Establishes that durations are estimates until the video is built and watched, and that edits must never be driven by target-length arithmetic. - Durable Documentation: Rules for writing notes, storyboards, and plans that a first-time reader can use, with intent at the head and dated measurements appended at the foot. - Use Case: While building a Remotion film, the assistant is asked to trim a scene to hit a 60-second target. Following this Skill, it reports the overrun instead of cutting, and lets the person decide. ## Quick Start Ask the assistant to review the current video project against the working-practice rules before setting any durations or calling the work finished.

Frequently Asked Questions about working-practice

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How should an AI assistant handle video duration estimates?▼

Treat durations as provisional until the video is built and watched, then measure against something playable like recorded narration. Never cut material to hit a target length; report the overrun and let the person decide.

What decisions belong to the human in AI-assisted video production?▼

The human judges meaning, humour, pacing, metaphor, and composition, since these are settled by watching. The assistant verifies checkable facts: rendering, resolution, safe areas, frame placement, and file consistency.

How do I write project notes that AI agents can use later?▼

State the rule and its reason together rather than recording a correction, since agents have no memory of the session. Define new terms where they first appear and append dated measurements at the foot of documents.

Why do AI assistants invent constraints during video work?▼

Assistants add unsourced rules while writing instructions, then defend them through repeated failures. When a second attempt fails the same way, list all constraints in play and refine only those the person actually set.

When should measurements go in a storyboard or plan document?▼

Fixed decisions like resolution, frame rate, and scenes belong at the head. Measured or revised figures are appended at the foot with dates, so the document does not imply the measurement preceded the work.