name: update-dip
description: Updates an author's Digital Identity Profile with new data from transcripts. Use when a DIP needs to be refreshed after a new transcript, feedback, or a scheduled review. IMPORTANT: the DIP is built ONLY from transcripts and recorded speech, NEVER from posts.
disable-model-invocation: true
argument-hint: [Author]
allowed-tools: Read, Grep, Glob, Edit
Update Digital Identity Profile
Refresh an author's DIP with new data from transcripts and recorded speech.
CRITICAL PRINCIPLE
The DIP is built ONLY from source material — transcripts, meeting recordings, talks, podcasts. NEVER from posts. Posts are OUTPUT of the DIP, not INPUT. Building the DIP from posts → feedback loop → loss of the real voice.
Input
$0 — author name (Seva, Kirill)- Conversation context — a new transcript, feedback, or a scheduled-update request
Required files
-
Current author DIP:
Knowledge-Base/01-Identity-Profiles/[Author]-DIP.md
-
DIP Update Workflow:
Templates/DIP-Update-Workflow.md
Data sources (hierarchy)
| Priority | Source | Path |
|---|
| 1 | Podcasts and talks | your-raw-sources/podcasts-and-events/ |
| 2 | Conferences | Media activities/Cursor Case Activities/Conferences/ |
| 3 | Livestreams | Media activities/Cursor Case Activities/Livestreams/ |
| 4 | AGI/Product calls | your-transcripts/internal-product-calls/ |
| 5 | Presales calls | Presales/Projects/Projects list/*/Transcripts/ |
| 6 | Marketing calls | your-transcripts/marketing-meetings/ |
| 7 | AI positioning calls | your-transcripts/strategy-calls/ |
| 8 | Client calls | Production/Projects/Projects list/*/Meetings transcripts/ |
| 9 | Granola syncs | your-transcripts/synced-meetings/ |
| 10 | Webinars | your-transcripts/webinars/ |
| 11 | Telegram (the author's own words) | your-raw-sources/telegram-posts/ |
NOT sources: LinkedIn posts (Posted/), PR comments, presentations, marketing copy, product blurbs.
Pipeline
Follow the process from _DIP-Update-Workflow.md.
Three update types:
1. New transcript / material:
- Read the new transcript — focus ONLY on the author's speech
- Extract: new stories, speech patterns, frameworks, contrarian positions, metrics, numbers
- Check for contradictions with the current DIP (the person may have changed their mind)
- Update the relevant DIP sections (Stories Bank, Frameworks, Expertise Clusters, Tone of Voice)
- Update the Source Index
- Add an entry to the Update Log
2. Feedback (from the author personally):
- Record it in the "FEEDBACK LOG & LEARNINGS" section
- Format: Date | Context | Category | Feedback | Lesson Learned
- Categories: tone | accuracy | context | angle | format | strategy
- If a pattern repeats 3+ times → promote it from the Feedback Log to the main DIP sections
3. Fresh Scan (before creating content):
- Check
Last Updated and Last Fresh Scan in the DIP
- Find ALL new transcripts featuring the author after that date
- Quickly scan each (5 min/file): new positions, stories, numbers, phrases
- If there are material updates → update the DIP before content
- Update the
Last Fresh Scan date
4. Quarterly review:
- Audit every section for accuracy
- Compare against the 2–3 most recent transcripts — does the voice still match?
- Verify: are there unprocessed transcripts?
- Update metrics (team size, clients, numbers)
- Increment the version in the Update Log
