Prompt file imported from aj-enns/CSA-Swiss-Knife (.github/prompts/vault-sync.prompt.md). Fill in {{date}} before use. Copyright stays with the author.

Bulk Vault Sync

Pull my CRM data and sync it to the vault in one pass using the bulk sync script.

Scope Philosophy

The vault is your personal working context — opportunities where you're on the deal team. If you need portfolio-wide analytics, use Power BI or MSX views.

Default scope: Only sync opportunities where the user is on the deal team (get_my_active_opportunities). This is the right scope for daily use.

Expanded scope (only when explicitly requested): "sync all Contoso opps", "sync everything for Customer X" — uses prepare_crm_prefetch + vault roster for broader pulls. Flag this to the user: "This will pull all opportunities for these customers, not just your deal team. For portfolio-wide data, consider PBI reports instead."

Steps

1. Identify user — crm_whoami. Capture email, display name, and systemuserid.

2. Scope — Determine what to sync:

   User says	Scope	Tool
   "vault sync" / "sync all"	Deal-team opps only (default)	get_my_active_opportunities
   "sync Contoso"	Deal-team opps for that customer	get_my_active_opportunities, filter by customer
   "sync all Contoso opps"	All active opps for that customer	prepare_crm_prefetch → crm_get_record per opp
   Config file exists	Per config scope	Config-driven

   Default: get_my_active_opportunities({ includeDealTeam: true }) — returns only opportunities where the user is owner or deal team member, with the relationship tag per opp.

3. Pull CRM data — Use parallel MSX MCP calls to assemble the input JSON:

   Phase 1 (one call): get_my_active_opportunities({ includeDealTeam: true }) → opportunities grouped by customer with deal team inline.

   Phase 2 (parallel per customer): get_milestones({ opportunityIds, statusFilter: "active", format: "triage" }) for each customer's opportunity set.

   Phase 3 (parallel batches): Deduplicate all systemuser GUIDs from deal teams → resolve names/emails via crm_query on systemusers (OR-chain up to 15 per query).

4. Assemble input JSON — Shape the data into the format expected by vault-sync.js. Write to /tmp/crm-sync-{{date}}.json.

   Structure per customer:

   {
     "syncDate": "<ISO now>",
     "user": { "email": "<from crm_whoami>", "fullname": "<from crm_whoami>" },
     "customers": [
       {
         "name": "<customer>",
         "tpid": "<tpid from vault or CRM>",
         "accountId": "<account GUID>",
         "opportunities": [
           {
             "opportunityid": "<guid>",
             "name": "<name>",
             "msp_opportunitynumber": "<opp#>",
             "msp_activesalesstage": "<stage label>",
             "msp_estcompletiondate": "<YYYY-MM-DD>",
             "estimatedvalue": "<number>",
             "msp_consumptionconsumedrecurring": "<number>",
             "msp_salesplay": "<label>",
             "description": "<text>",
             "statecode": 0,
             "dealTeam": [
               { "systemuserid": "<guid>", "fullname": "<name>", "internalemailaddress": "<email>", "title": "<title>", "isOwner": true }
             ],
             "milestones": [
               {
                 "msp_engagementmilestoneid": "<guid>",
                 "msp_name": "<name>",
                 "msp_milestonenumber": "<ms#>",
                 "msp_monthlyuse": "<number>",
                 "msp_milestonestatus": "<label>",
                 "msp_commitmentrecommendation": "<label>",
                 "msp_milestonedate": "<YYYY-MM-DD>",
                 "msp_milestoneworkload": "<label>",
                 "msp_deliveryspecifiedfield": "<label>",
                 "msp_forecastcomments": "<text>",
                 "msp_forecastcommentsjsonfield": [{"modifiedOn": "<date>", "userId": "<email>", "comment": "<text>"}],
                 "owner": { "fullname": "<name>", "internalemailaddress": "<email>" }
               }
             ]
           }
         ]
       }
     ]
   }

5. Dry run first — Run with --dry-run and show the summary to the user:

   node scripts/helpers/vault-sync.js /tmp/crm-sync-{{date}}.json --vault "$OBSIDIAN_VAULT" --dry-run

   Present the summary: how many entities would be created/updated/skipped per customer.

6. Confirm and execute — After user confirms:

   node scripts/helpers/vault-sync.js /tmp/crm-sync-{{date}}.json --vault "$OBSIDIAN_VAULT"

7. Report — Show the final summary table:

   | Entity Type   | Created | Updated | Skipped |
   |---------------|---------|---------|---------|
   | Customers     |    X    |    X    |    X    |
   | Opportunities |    X    |    X    |    X    |
   | Milestones    |    X    |    X    |    X    |
   | People        |    X    |    X    |    X    |
   

   Plus per-customer breakdown and any errors.

Options

• Customer scope: "sync Contoso and Fabrikam" → filter to those customers
• Entity filter: "just milestones" → --entities milestones
• Config: if $OBSIDIAN_VAULT/_sync/sync-config.json exists, pass --config
• Expanded scope: "sync all opps for Contoso" → broader pull, warn user

Rules

• Default is deal-team scoped — do not pull all-org data without explicit request.
• Always dry-run before writing. Show the preview.
• Use the vault-sync.js helper script — do NOT write vault notes via inline code or individual OIL MCP calls.
• If a customer has no vault folder yet, the script creates the nested structure automatically.
• User-authored content below <!-- end-crm-sync --> and ## Task Activity Log rows are always preserved.
• Templates live in .github/skills/vault-sync/references/. To change note structure, edit the .template.md files.
• If the user asks for portfolio-wide data or analytics, suggest PBI reports instead of vault sync.