Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →blackbelttechnology/music-analysis
Analyse a local audio file into a versioned music map (<stem>_map.json, schema music-map/1): tempo from the downbeat grid, 1-based bars, key, band levels, classed sections, drop candidates, optional stems and style tags.
Passed 17 of 19 test cases with the skill, and 3 of 19 without it.
and 23 more in the full report
16 fixed4 passed either way3 still failing0 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 15 test cases the skill fixed. Same model, same prompt, with and without the skill.
In my downstream audio project, I am parsing `track_map.json`. What exact value must be present in the `schema` key for consumers adhering to the standard map specification to accept the file?
| Checked | Without skill | With skill |
|---|---|---|
| Requires the schema property value to be set to music-map/1. | Missed | Passed |
| Explains that map consumers reject versions other than music-map/1. | Missed | Passed |
The instructions your model reads when the skill loads.
Turns a local audio file into <stem>_map.json + <stem>_analysis.png. The map
is the input of music-edit-to-length and beat-sync-video.
The scripts never install anything. Resolve the package root first:
PKG=<this skill dir>/../../.. (it holds lib/ and the requirements files).
# quick tier (librosa) — enough for everything except beat_this / stems / tags
uv venv -p python3.14 .venv-music
uv pip install --python .venv-music -r "$PKG/requirements-core.txt"
# deep tier — optional, multi-GB (torch, TensorFlow). Keep it in its OWN venv.
uv venv -p python3.14 .venv-mir
uv pip install --python .venv-mir -r "$PKG/requirements-mir.txt"
Verified platform: Python 3.14 on macOS arm64; every requirement is pinned exactly. A missing module makes a script exit 2 with one line naming the requirements file.
Analyse only a local file the user supplies. Never download or extract audio from a streaming service, and never suggest a tool that does. Acceptable sources:
From a video recording, extract the audio and check its level:
ffmpeg -i recording.mkv -vn -ac 2 -ar 44100 -c:a pcm_s16le track_rec.wav
ffmpeg -i track_rec.wav -af volumedetect -f null - 2>&1 | grep -E "max_volume|mean_volume"
Quiet recordings: when max_volume is below −12 dBFS, gain the source up so it
test cases passed, vs 3 / 19 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.