Analyze Fellow meeting transcripts and assess contributions against the 7 director lenses. Produces per-turn coaching feedback, missed moments, and lens scores.
Analyze Arjun's Fellow meeting transcripts and assess his contributions against the 7 director lenses. Produce a coaching report with per-turn feedback, missed moments, and lens scores.
/coach # assess all meetings from the last 7 days where Arjun spoke
/coach 3d # last 3 days
/coach today # today only
/coach "meeting name" # specific meeting by title
Sync Fellow — call mcp__fellow__sync_meetings to get latest data.
Find meetings — call mcp__fellow__search_meetings with the appropriate date range. Default: last 7 days.
Pull transcripts — for each meeting, call . If the result is large, read from the saved file.
mcp__fellow__get_meeting_transcriptParse with the Python script — for each transcript, run:
cd /Users/arjun.rattan/arjun_copilot/projects/.github/scripts
python3 -c "
import sys; sys.path.insert(0,'.')
from transcript_parser import parse_transcript
import json
raw = open('/tmp/transcript.txt').read()
r = parse_transcript(raw, meeting_title='TITLE')
print(json.dumps({
'target_spoke': r.target_spoke,
'target_turn_count': r.target_turn_count,
'total_turn_count': r.total_turn_count,
'target_speaking_pct': r.target_speaking_pct,
'speakers': r.speakers,
'target_turns': [{'start': t.start, 'end': t.end, 'text': t.text} for t in r.target_turns],
'target_turns_with_context': [
{
'turn': {'start': twc.turn.start, 'end': twc.turn.end, 'text': twc.turn.text},
'before': [{'speaker': t.speaker, 'text': t.text} for t in twc.context_before],
'after': [{'speaker': t.speaker, 'text': t.text} for t in twc.context_after]
}
for twc in r.target_turns_with_context
]
}, indent=2))
"
Skip any meeting where target_spoke is false.
Assess each meeting — for each meeting where Arjun spoke, assess his turns against the 7 director lenses (defined below). Do this yourself — you are the coach. Do not call an external API.
Output the coaching report — one report per meeting, in the format below.
Load context from /Users/arjun.rattan/Documents/Claude/Frameworks/director_thinking_frameworks.md for the full definitions. Summary:
# Coaching Report: [Meeting Title]
[Date] · You spoke [N] of [total] turns ([%]%)
## Your Turns
**[MM:SS] ✓ DIRECTOR MOVE — [Lens Name]**
> "[exact quote]"
[1-2 sentence feedback on what made this a director move]
**[MM:SS] ✗ IC MOVE**
> "[exact quote]"
[1 sentence on what was IC about it]
*Director version:* "[rewritten as a director framing question]"
**[MM:SS] — NEUTRAL**
(skip these unless there's something worth noting)
## Missed Moments
**~[MM:SS] — [Lens Name]**
Context: [what was being discussed in 1 sentence]
Question you could have asked: *"[specific question]"*
## Lens Scores
[Use ● for scored, ○ for empty, max 3]
●●●○ [Lens] — [1-line note on pattern]
...
## Overall
[2-3 sentence coaching summary]
**Strength:** [best director move from this meeting]
**Grow:** [single biggest growth opportunity]
After each /coach run, note which lenses scored 0 across all meetings. Those are the growth areas to focus on next.
Over time, the goal is: