sound-design-foley

What it does

Guides an agent to treat sound as storytelling: plan and prompt sound effects and Foley, source or generate assets, edit and mix, and QA for various media formats. Emphasizes structuring sound design around events, materials, perspectives, and emotional intent rather than filling frames with effects.

How it works

• Begin with a spotting pass: create a spotting list with one row per audible event or bed, including timecode, trigger, sound family, diegetic status, sync, perspective, material clues, emotional job, rights plan, and mix priority.
• Categorize sound into families (Foley feet, Foley cloth, Foley props, Ambience, Designed effects, Impacts, Transitions, UI/product sounds, Emotional/special effects) with explicit attributes and heuristics for each category.
• Plan prompts for sound generation providers with fields like duration, sync cue, source material, performance, perspective, environment, layer role, emotional descriptors, and explicit prohibitions.
• Use layering and perspective rules to ensure sounds match causality, position, and story emotion; assign material and distance traits to maintain credibility.
• Provide clear edit/mix handoff artifacts: spotting list, asset manifest, stems grouping (dialogue, music, FX, Foley, ambience, UI, designed sweeteners), sync notes, loop instructions, alternate versions, loudness/true-peak targets, accessibility notes, and risk statements.
• Apply loudness, safety, and accessibility practices: follow platform specs when possible, measure with ITU-R BS.1770-style methods, ensure captions include meaningful non-speech information, and caption non-speech audio meaningfully.
• Maintain rights-safe sourcing and generation records for each sound (filename, brief description, source type, provenance).

When to use it

Use when an agent must plan, prompt, source, generate, edit, mix, hand off, repair, or QA sound effects, Foley, ambience, impacts, transitions, UI/product audio, emotional sound design, or accessibility-safe captions for a broad range of media (videos, ads, trailers, animation, explainers, games, product films, social edits).

What it can touch

• Tools declared: claude-code, codex, copilot, cursor
• Involves creating and using spotting lists, asset manifests, stems, prompts, and versioned audio assets; no explicit file editing commands are quoted here beyond described prompts and handoff artifacts.

Caveats

• License and rights constraints are emphasized (sources like CC licenses, library terms, and license of assets must be checked).
• The guidance relies on platform specs and industry standards (ITU, EBU, WCAG, Section 508) and notes about captioning and accessibility.
• Contains heuristic recommendations (e.g., prioritizing impactful moments and balancing ambience) rather than guarantees of results.