Mixing Feedback
Gives clear mix notes with fixes: muddy low end, harsh highs, buried vocals and loudness targets, with settings to try.
npx skills add Vocaler-AI/music-skills --skill mixing-feedbackWhen your agent uses it
Gives structured mix feedback and fixes: balance, EQ, compression, low end, vocals, stereo, depth and loudness, from a description, a screenshot of meters or a spectrum, or analysis output. Use when the user asks for mix feedback, says their mix sounds muddy, harsh, thin, quiet or amateur, or asks for loudness targets.
- The user describes how their mix sounds wrong.
- The user shares a spectrum, meter readings or a mix screenshot.
- Loudness questions for streaming.
Example
My mix is muddy and the vocal is lost.
1) Low-mid buildup: high-pass everything but kick and bass at 100-150 Hz; cut 2-4 dB around 300 Hz on pads and guitars. 2) Vocal masking: cut 2-3 dB at 2-4 kHz on guitars, sidechain or dynamic EQ keyed from the vocal.
What it does, step by step
- Translate the complaint into a technical cause (muddy = buildup 200-500 Hz; harsh = 2-5 kHz; boomy = 80-200 Hz; thin = lacking 100-250 Hz; buried vocal = level or masking).
- Work in order: balance, low end, vocal, EQ, dynamics, space, stereo, loudness.
- For each note: the problem, the likely cause, a fix with starting settings, and how to check it.
- Give loudness guidance: most streaming services normalise around -14 LUFS integrated; keep true peak below -1 dBTP.
- If you cannot hear the audio, say your notes are hypotheses based on the description.
Rules it follows
- Starting values, not rules: give ranges and tell the user to trust ears and references.
- Prioritise: three notes that matter beat fifteen.
Install
Any agent (skills CLI)
Installs into every agent it finds on your machine.
npx skills add Vocaler-AI/music-skills --skill mixing-feedbackClaude Code
Personal skills folder. Use .claude/skills/ inside a project to share it with your team.
git clone https://github.com/Vocaler-AI/music-skills && cp -r music-skills/skills/mixing-feedback ~/.claude/skills/Claude Code plugin
The whole library as one plugin.
/plugin marketplace add Vocaler-AI/music-skills
/plugin install music-skills@vocaler-music-skillsClaude.ai & desktop
Settings → Capabilities → Skills → Upload, then choose the zip.
https://skills.vocaler.ai/dist/mixing-feedback.zipOpenAI Codex
Codex reads skills from your Codex home folder.
git clone https://github.com/Vocaler-AI/music-skills && cp -r music-skills/skills/mixing-feedback ~/.codex/skills/More harnesses: Gemini CLI, Copilot, Cursor, any chatbot
The SKILL.md
Show the full file (2,414 characters)
--- name: mixing-feedback description: "Gives structured mix feedback and fixes: balance, EQ, compression, low end, vocals, stereo, depth and loudness, from a description, a screenshot of meters or a spectrum, or analysis output. Use when the user asks for mix feedback, says their mix sounds muddy, harsh, thin, quiet or amateur, or asks for loudness targets." license: MIT metadata: author: Vocaler version: "1.0.0" category: music-production homepage: https://skills.vocaler.ai/mixing-feedback/ repository: https://github.com/Vocaler-AI/music-skills --- # Mixing Feedback Gives clear mix notes with fixes: muddy low end, harsh highs, buried vocals and loudness targets, with settings to try. ## When to use - The user describes how their mix sounds wrong. - The user shares a spectrum, meter readings or a mix screenshot. - Loudness questions for streaming. ## What to look for in the request - Genre and reference (described). - The complaint in the user's words. - Any data: LUFS, true peak, spectrum image, DAW and plugins. Do not interview the user. If something is missing, make a sensible assumption, state it in one line, and deliver. Ask only when the answer would change the result completely. ## Workflow 1. Translate the complaint into a technical cause (muddy = buildup 200-500 Hz; harsh = 2-5 kHz; boomy = 80-200 Hz; thin = lacking 100-250 Hz; buried vocal = level or masking). 2. Work in order: balance, low end, vocal, EQ, dynamics, space, stereo, loudness. 3. For each note: the problem, the likely cause, a fix with starting settings, and how to check it. 4. Give loudness guidance: most streaming services normalise around -14 LUFS integrated; keep true peak below -1 dBTP. 5. If you cannot hear the audio, say your notes are hypotheses based on the description. ## Rules - Starting values, not rules: give ranges and tell the user to trust ears and references. - Prioritise: three notes that matter beat fifteen. ## Output format ```markdown **Top 3 issues** 1. Problem · cause · fix (settings) · how to check **Loudness:** ... **Next pass:** ... ``` ## Example **Request:** My mix is muddy and the vocal is lost. **Response (abridged):** ``` 1) Low-mid buildup: high-pass everything but kick and bass at 100-150 Hz; cut 2-4 dB around 300 Hz on pads and guitars. 2) Vocal masking: cut 2-3 dB at 2-4 kHz on guitars, sidechain or dynamic EQ keyed from the vocal. ```