wrap-up

name	wrap-up
description	Format and upload session wrap-up to NotebookLM for persistent semantic memory
version	1.0.0
author	Trinity S3AI Framework

Wrap-Up Skill

Upload session summaries to NotebookLM for cross-session memory persistence.

What It Does

1. Extracts session context from .trinity/ state files
2. Formats summary as Markdown with metadata
3. Uploads to NotebookLM as searchable source

Usage

/wrap-up "Session completed Phi Loop iterations for Ring-071"

Or with full details:

/wrap-up --summary "Implemented NotebookLM backend" \
         --decisions "Used notebooklm-py SDK with cookie auth" \
         --files "contrib/backend/notebooklm/*.py" \
         --steps "Run integration tests"

Implementation

This skill uses the t27 spec-first approach:

• Spec: specs/automation/wrapup-auto.t27
• Backend: contrib/backend/notebooklm/wrapup_auto.py
• Invocation: Python script directly (no shell scripts - L7 compliant)

Direct invocation (for debugging):

.trinity/notebooklm-venv/bin/python3 \
    contrib/backend/notebooklm/wrapup_auto.py \
    --summary "Session summary" \
    --session-id "$(git rev-parse --short HEAD)" \
    --decisions "Key decisions" \
    --files "file1.py,file2.py" \
    --steps "Next steps"

Configuration

• Auth: Cookie-based via notebooklm login (stores in ~/.notebooklm/storage_state.json)
• Active Notebook: Set via --notebook flag (default: "t27-QUEEN-BRAIN")
• Default Notebook: "t27-QUEEN-BRAIN" (creates if not exists)
• Storage: ~/.notebooklm/ — browser profile, storage state

Setup Commands:

python -m venv .trinity/notebooklm-venv
.trinity/notebooklm-venv/bin/pip install notebooklm-py
notebooklm login              # Authenticate via browser (one-time)

Output

Returns source ID and notebook ID for verification:

{
  "notebook_id": "...",
  "source_id": "...",
  "uploaded_at": "2026-04-08T..."
}