analyze-meetings

Context

Arguments: $ARGUMENTS

---

Instructions

Step 1: Resolve the Meetings Path

• If $ARGUMENTS is provided, use it as the meetings folder path
  • Expand ~ if needed: bash -c 'echo <path>'
  • Verify the folder exists and has readable files: ls "<resolved-path>"
  • If the folder doesn't exist, tell the user and stop
• If no path was provided, ask: "Please provide the path to your meetings folder (e.g. ~/meetings/)."

Step 2: Discover Available Data

Scan the folder for transcript files (.txt, .md, .vtt, .srt, .docx):

ls "<meetings-path>"

• Confirm how many files and what date range they cover
• Check whether files contain speaker labels (e.g. Sam:, [Sam], Speaker 1:)
• Identify the user's name or label in the transcripts — ask if unclear

If no transcript files are found, tell the user what formats are supported and stop.

Step 3: Clarify Analysis Goals

If the user hasn't specified what to look for, ask what they want to learn:

• Specific behaviors: conflict avoidance, interruptions, filler words
• Communication effectiveness: clarity, directness, listening
• Meeting facilitation and leadership style
• Speaking ratios and turn-taking patterns
• Tracking improvement over time

Step 4: Read and Analyze Transcripts

Read each relevant transcript file. For each requested analysis dimension:

Conflict Avoidance

• Hedging language: "maybe", "kind of", "I think", "sort of", "potentially"
• Indirect asks instead of direct requests
• Changing subject when tension arises
• Non-committal agreement: "yeah, but...", "whatever you think"
• Not naming obvious problems when they come up

Speaking Ratios

• Estimate percentage of each meeting the user is speaking
• Count interruptions (both by and of the user)
• Note question vs. statement ratio
• Track average speaking turn length

Filler Words

• Count: "um", "uh", "like", "you know", "actually", "basically", etc.
• Frequency per speaking turn, and in which situations they spike