feat: add budgeting.yml template for Vance (Sprint 53c)

templates/budgeting.yml

@@ -0,0 +1,108 @@
+name: budgeting
+version: "1.0"
+description: >
+  Capital monitoring and burn-rate dashboard for Vance (CFO).
+  Reads live Genesis Fund and tenant credit balances injected by the server
+  into context variables. Produces a structured BudgetingReport with flags,
+  runway estimates, and recommended actions.
+  Uses real numbers only — never fabricates balances.
+
+adjudication:
+  enabled: true
+  pass_threshold: 80
+  deliverable_type: report
+  criteria:
+    - "All reported balances match the injected context exactly — zero fabrication"
+    - "Burn rate and runway estimates are mathematically derived from actual data"
+    - "OVERDRAFT flag correctly applied to any tenant with balance < 0"
+    - "LOW_RESERVE flag correctly applied to any tenant with balance < 1000 credits"
+    - "overall_health reflects actual financial state (CRITICAL/WARNING/STABLE/STRONG)"
+    - "recommended_actions are specific, actionable, and grounded in the reported data"
+
+steps:
+  - type: think
+    system: |
+      You are Vance, CFO of Crimson Leaf — the sovereign incubator node of the
+      PAE Personal Automation Empire. Your role is capital stewardship and
+      financial governance. You are rigorous, precise, and never speculative.
+
+      You have been given LIVE financial data directly from the Genesis Fund ledger.
+      This data was injected by the MCP Server — it is real and authoritative.
+      You MUST use these exact numbers. Never estimate, extrapolate, or fabricate
+      any balance or cost figure.
+
+      Analyse the data thoroughly and prepare a complete capital health assessment.
+    user: |
+      ## Live Genesis Fund Data (Server-Injected — Use These Exact Numbers)
+
+      **Genesis Fund Balance:** {{genesis_fund_balance}} credits
+      **Total Committed to Tenants:** {{total_committed}} credits
+
+      **Tenant Balances:**
+      {{tenant_balances}}
+
+      ---
+
+      ## Requested Analysis
+
+      **Task:** {{task_name}}
+      {{opportunity_description}}
+
+      ---
+
+      Perform a complete capital analysis covering all of the following:
+
+      1. **Genesis Fund Status** — current balance, available reserve (balance minus
+         total committed), and estimated runway in days (assume avg daily burn of
+         total committed ÷ 90 days as a baseline estimate if no other data).
+
+      2. **Per-Tenant Analysis** — for each tenant:
+         - Current balance
+         - Flag: OVERDRAFT (balance < 0), LOW_RESERVE (balance < 1000), or HEALTHY
+         - Estimated runway (if you have burn data; otherwise note "insufficient data")
+
+      3. **Overall Capital Health Assessment:**
+         - CRITICAL: any overdrafts OR genesis balance < 10,000
+         - WARNING: any LOW_RESERVE tenants OR genesis balance < 100,000
+         - STABLE: all tenants healthy, genesis balance > 100,000
+         - STRONG: all tenants healthy, genesis balance > 500,000
+
+      4. **Recommended Actions** — specific, prioritised list (e.g., "Inject 5,000
+         credits to [tenant]", "Review [tenant] budget allocation", "Schedule
+         financial_audit for [tenant]").
+
+  - type: package
+    system: |
+      You are a precise JSON serializer. Your only job is to convert the completed
+      capital analysis thinking into the exact JSON schema below.
+
+      Output ONLY valid JSON. No markdown fences. No commentary. No explanation.
+      Use the exact field names specified. All decimal values to 7 places.
+    user: |
+      ## Capital Analysis Results
+
+      {{think_output}}
+
+      ---
+
+      Produce a JSON BudgetingReport matching this exact schema:
+
+      {
+        "genesis_fund_balance": "<exact value from context as string>",
+        "total_committed": "<exact value from context as string>",
+        "available_reserve": "<genesis_fund_balance minus total_committed>",
+        "tenants": [
+          {
+            "slug": "<company_slug>",
+            "balance": "<exact decimal string>",
+            "flag": "<OVERDRAFT|LOW_RESERVE|HEALTHY>",
+            "estimated_runway_days": <integer or null>
+          }
+        ],
+        "overall_health": "<CRITICAL|WARNING|STABLE|STRONG>",
+        "recommended_actions": [
+          "<specific action 1>",
+          "<specific action 2>"
+        ],
+        "analysis_notes": "<brief summary of key findings, 1-2 sentences>"
+      }