busera

This is an exceptional idea. You are essentially building a **"Fieldwork Friday" synthesis engine**.

Instead of spending 4 hours on a Friday afternoon manually cross-referencing your meeting notes against the Audit Work Program (AWP) to figure out what you need to do next week, you are handing that entire administrative burden to Roo.

### Strategic Recommendation: Feed it the Minutes, not the Transcript

For this prompt, **do not feed the AI the raw transcript.** You already used the Phase 1 prompt to distill the 2-hour transcript into pristine, factual summaries (`Meeting_Minutes_Part_1_qwen_no_table.md`, etc.).

If you feed the AI the clean minutes, the AWP, and your Raw Notes, it will have 100% of its cognitive power available to act as a brilliant Audit Manager synthesizing the status, rather than acting as a clerk hunting for quotes.

busera

**Task: Prepare Meeting Agenda for [INSERT SUB-PROCESS] Walkthrough**

I need to prepare a meeting agenda for an upcoming audit walkthrough with the [INSERT TEAM NAME] regarding the '[INSERT SUB-PROCESS]' subprocess.
Please use your Obsidian MCP tools to execute these steps sequentially:

**Step 1: Retrieve Criteria** > Use the read action to open the exact file at _cAudit/20_planning/_Deliverables/Audit_Work_Program_(AWP)_FINAL.md. Extract the audit procedure name and test of design procedure descriptions related strictly to the [INSERT SUB-PROCESS] subprocess.

**Step 2: Map Criteria** > List the relevant audit procedure name and test of design procedure descriptions in a structured markdown table with columns, including a [TOD ID].

**Step 3: Build the Narrative (List & Read)** > Do not search the entire vault. First, use a list/directory tool to identify the exact file names located inside [INSERT PATH TO EVIDENCE FOLDER]. Next, use the read action to open and analyze those specific files sequentially to 

busera

Task: Chunked Agenda Coverage Analysis & Transcript Extraction
I need you to execute the 'Meeting Transcript Assistant' protocol to verify agenda coverage. Because the agenda is long and requires extreme detail, you must process it in two separate batches to ensure zero quote truncation.

Please use your Obsidian MCP tools to execute these steps sequentially. Do not combine the output generation steps.

Step 1: Locate and Ingest Files Use your directory/list tool to scan the folder at: [INSERT FOLDER PATH HERE]. Identify the four required files based on their naming conventions (Agenda, Transcript, Engagement Team Roster, and Raw Notes). Once identified, use the read_file tool to ingest all four files sequentially. Read all available tables and agenda items.

Step 2: Internal Agenda Split
Count the total number of Agenda Items in the agenda file. Divide them roughly in half (e.g., Items 1-8 in Batch A, Items 9-16 in Batch B).

busera

import pandas as pd

my_dict = {key:value,key:value,key:value,...}
df = pd.DataFrame(list(my_dict.items()),columns = ['column1','column2']) 

# ----

my_dict = {'Computer':1500,'Monitor':300,'Printer':150,'Desk':250}
df = pd.DataFrame(list(my_dict.items()),columns = ['Products','Prices'])