Minutes setup

Skill silverstein/minutes/tooling/skills/goldens/codex/minutes-setup

Guided first-time setup for Minutes — download whisper model, create directories, configure audio input. Use when the user says "set up minutes", "install minutes", "first time setup", "configure minutes", "get started with minutes", "how do I start using minutes", or when verify shows missing components.From its SKILL.md

Install

npx -y skills add silverstein/minutes --skill minutes-setup

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

SKILL.md

2.9 KB, 701 tokens by cl100k_base, as published. Nobody here has run it

Skill Path

Before running helper scripts or opening bundled references, set:

export MINUTES_SKILLS_ROOT="$(git rev-parse --show-toplevel)/.agents/skills/minutes"
export MINUTES_SKILL_ROOT="$MINUTES_SKILLS_ROOT/minutes-setup"

/minutes-setup

Walk the user through first-time Minutes setup, step by step.

Setup steps

1. Check current state first

Run the verify skill's script to see what's already done:

bash "$MINUTES_SKILLS_ROOT/minutes-verify/scripts/verify-setup.sh"

Skip any steps that already pass.

2. Build the binary (if needed)

cd ~/Sites/minutes
export CXXFLAGS="-I$(xcrun --show-sdk-path)/usr/include/c++/v1"
cargo build --release

The binary lands at target/release/minutes. The user should add it to their PATH or create a symlink.

3. Download a whisper model

Ask the user which quality level they want using AskUserQuestion:

Model	Size	Speed	Quality	Best for
`tiny`	75 MB	~10x real-time	Low	Quick tests, short memos
`small`	466 MB	~4x real-time	Good	Daily meetings (recommended)
`medium`	1.5 GB	~2x real-time	Great	Important meetings, accents
`large-v3`	3.1 GB	~1x real-time	Best	Legal, medical, foreign language

Then run:

minutes setup --model <chosen-model>

4. Create directories

mkdir -p ~/meetings/memos

5. Audio input (if recording calls)

For in-person conversations, the built-in mic works fine. For Zoom/Meet/Teams:

Install BlackHole: brew install blackhole-2ch
Open Audio MIDI Setup (Spotlight → "Audio MIDI Setup")
Create a Multi-Output Device combining speakers + BlackHole
Set the Multi-Output Device as system output
Set BlackHole as Minutes' input (or system default input)

See minutes-record/references/audio-devices.md for the full guide.

6. Verify

Run verify again to confirm everything passes:

bash "$MINUTES_SKILLS_ROOT/minutes-verify/scripts/verify-setup.sh"

7. Test recording

minutes record --title "Test recording"
# Speak for 10-15 seconds
minutes stop

Check the output file exists in ~/meetings/ and has a transcript.

Gotchas

macOS 26 (Tahoe) requires CXXFLAGS — The whisper.cpp build needs the C++ include path set explicitly. This is a known Apple SDK issue.
First model download can be slow — The small model is 466 MB. On slow connections, tiny is a good starting point (75 MB).
BlackHole setup is the hardest part — Most users struggle with the Audio MIDI Setup step. Offer to walk through it if they get stuck.

What ships with it: 1 file

226 B alongside SKILL.md

agents/

openai.yaml226 B

Gives 2 of the 12 instructions most project setup skills give in 701 tokens

Counted across 1,553 of the 3,091 authors here whose files we hold, read 2026-09-06

Write the configuration filein 36 of 1553
Create the directory structurehere, and in 35 of 1553, across 33 files
Verify the setuphere, and in 31 of 1553, across 28 files
Run the setup scriptin 30 of 1553, across 29 files
Pre-determine the required sample sizein 29 of 1553, across 12 files
Check if the configuration already existsin 29 of 1553
Document every testin 26 of 1553, across 10 files
Start with a hypothesisin 26 of 1553, across 11 files
Ask one question at a timein 22 of 1553
Test a single variable per testin 21 of 1553, across 9 files
Read product marketing context before asking questionsin 19 of 1553, across 8 files
Do not peek and stop earlyin 18 of 1553, across 7 files

Said here and by no other author read

Download a whisper model
Configure audio input
Test recording

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.