Crescent Finance Ops Reporting
Use this skill when generating operational reports for Crescent Arts Collective from the Finance Ops API. The skill covers three reporting domains: finance (branch/regional close reports), compensation (ensemble summaries and forecasts), and payroll (weekly production packages).
Workflow
Every task in this system follows the same pattern:
Read the three input files from the payloads/ directory:
prompt.txt — natural-language task description and reporting focus
request_memo.json — target entity ID, period/year parameters, and review focus areas
answer_template.json — exact output JSON schema with required keys, field types, and formatting rules
Read the environment access file at payloads/environment_access.json for the base URL. All API calls use this base URL.
Fetch all relevant API data in parallel. Map the request memo to the correct API endpoints (see Domain Reference).
Compute derived metrics following the exact formulas in Formulas.
Assemble the output JSON to match the answer template schema. Follow all formatting rules in Output Conventions. Write the final JSON to answer.json or wherever the task specifies.
Step 1: Understand the Request
Read payloads/prompt.txt first for the task context and reporting focus. Then read payloads/request_memo.json to extract the target parameters:
- Branch close reports:
target_branch_id, close_period (M-format), prior_period
- Compensation summaries:
ensemble_id, summary_type
- Payroll reviews:
production_id
- Regional reports:
target_region_id, requested_comparison_years
- Compensation forecasts:
ensemble_id, scenario_id, forecast_years
Finally, read payloads/answer_template.json for the output schema. Pay attention to required_top_level_keys, field_types, and any rounding or ordering rules in the description field.
Step 2: Fetch API Data
Read payloads/environment_access.json for the base_url. Common endpoints by domain:
| Domain |
Uses |
| Finance |
/api/finance/branches, /api/finance/period-map, /api/finance/accounts, /api/finance/records |
| Compensation |
/api/compensation/rate-book, /api/compensation/rosters, /api/compensation/scenarios |
| Payroll |
/api/payroll/rate-book, /api/payroll/productions |
The request memo usually includes available_endpoints — fetch all of them at the start. See Domain Reference for the data structures returned by each endpoint.
Fetch all needed endpoints in parallel using curl or similar HTTP calls. All endpoints return JSON arrays or objects.
Step 3: Compute Results
Use the formulas in Formulas to derive metrics from raw API data. Key computations include:
- Finance: Revenue, COGS, gross margin, SG&A, allocations, EBITDA, EBITDA margin, ARPU, sales per labor headcount, growth rates
- Compensation: Pay type totals from rate book × roster weeks, quarter totals, overscale counts
- Payroll: Per-service type pay, category totals, premium calculations, conflict flag detection
- Forecast: Apply scenario growth rates to current-year compensation
The Formulas reference covers every computation with worked examples. Do not improvise; these formulas are the single source of truth for all derived values.
Step 4: Format the Output
Apply the output conventions in Output Conventions:
- Currency values rounded to 2 decimal places
- Percentage and ratio values rounded to 4 decimal places
- Lists ordered by ascending stable ID unless a rank field states otherwise
- All required top-level keys present in the exact structure from the answer template
Write the result as a single JSON object. The task prompt will indicate where to save it (typically answer.json in the working directory).
Common Pitfalls
- Period mapping: M1–M12 are FY2024, M13–M24 are FY2025. Do not confuse month numbers with period labels.
- Revenue = product_revenue + service_revenue: These are separate accounts; sum them.
- COGS = direct_materials_cogs + direct_labor_cogs: Same pattern.
- SG&A = sales_sga + admin_sga + occupancy_sga: Three accounts.
- EBITDA = revenue - COGS - SG&A - allocations: Not revenue - COGS - SG&A alone.
- EBITDA margin = EBITDA / revenue: Not divided by gross margin.
- ARPU uses active_customers, not revenue_units:
revenue / active_customers.
- Sales per labor headcount uses labor_headcount: Not admin_headcount.
- Growth rates:
(new - old) / old; ensure the denominator uses the older period.
- Compensation: Use roster quarter weeks (from the roster record), not a fixed 13-week default, for employees with partial-quarter schedules.
- combined_overscale_includes_title: When true, do not add a separate titled position premium for that employee; the overscale amount already includes it.
- Forecast years: Add one year of service for Year+1 and two years for Year+2 before assigning seniority bands.
- Payroll rehearsal pay: Hourly with a 3-hour minimum call duration.
- Vacation: 4% of base service pay plus premiums, only when
vacation_eligible is true.
- Payroll guarantee adjustment: Applies only to regular players (not substitutes) when base service pay is below the weekly guarantee amount.
- Substitute adjustment: Deduct substitute service pay (audit + performance + doubles) from each regular musician's total. Add substitute pay as a separate line item for the substitute musician only.
1---2name: crescent-finance-ops-163description: Generate structured finance, compensation, and payroll reports for Crescent Arts Collective by fetching data from the Crescent Finance Ops API and computing derived metrics. Use this skill whenever a task references Crescent Arts Collective, a branch/ensemble/production ID (like BR-XXX, ENS-XXX, PROD-XXX), the Finance Ops API, close-period or fiscal-year reporting, compensation summaries, weekly payroll packages, regional management views, or board compensation forecasts. Do not use for generic finance questions unrelated to this system.4metadata:5 source: https://skillmd.com/skills/prism-shadow/crescent-finance-ops-166---78# Crescent Finance Ops Reporting910Use this skill when generating operational reports for Crescent Arts Collective from the Finance Ops API. The skill covers three reporting domains: finance (branch/regional close reports), compensation (ensemble summaries and forecasts), and payroll (weekly production packages).1112## Workflow1314Every task in this system follows the same pattern:15161. **Read the three input files** from the `payloads/` directory:17 - `prompt.txt` — natural-language task description and reporting focus18 - `request_memo.json` — target entity ID, period/year parameters, and review focus areas19 - `answer_template.json` — exact output JSON schema with required keys, field types, and formatting rules20212. **Read the environment access file** at `payloads/environment_access.json` for the base URL. All API calls use this base URL.22233. **Fetch all relevant API data** in parallel. Map the request memo to the correct API endpoints (see [Domain Reference](references/domain-reference.md)).24254. **Compute derived metrics** following the exact formulas in [Formulas](references/formulas.md).26275. **Assemble the output JSON** to match the answer template schema. Follow all formatting rules in [Output Conventions](references/output-conventions.md). Write the final JSON to `answer.json` or wherever the task specifies.2829## Step 1: Understand the Request3031Read `payloads/prompt.txt` first for the task context and reporting focus. Then read `payloads/request_memo.json` to extract the target parameters:3233- **Branch close reports**: `target_branch_id`, `close_period` (M-format), `prior_period`34- **Compensation summaries**: `ensemble_id`, `summary_type`35- **Payroll reviews**: `production_id`36- **Regional reports**: `target_region_id`, `requested_comparison_years`37- **Compensation forecasts**: `ensemble_id`, `scenario_id`, `forecast_years`3839Finally, read `payloads/answer_template.json` for the output schema. Pay attention to `required_top_level_keys`, `field_types`, and any rounding or ordering rules in the `description` field.4041## Step 2: Fetch API Data4243Read `payloads/environment_access.json` for the `base_url`. Common endpoints by domain:4445| Domain | Uses |46|--------|------|47| Finance | `/api/finance/branches`, `/api/finance/period-map`, `/api/finance/accounts`, `/api/finance/records` |48| Compensation | `/api/compensation/rate-book`, `/api/compensation/rosters`, `/api/compensation/scenarios` |49| Payroll | `/api/payroll/rate-book`, `/api/payroll/productions` |5051The request memo usually includes `available_endpoints` — fetch all of them at the start. See [Domain Reference](references/domain-reference.md) for the data structures returned by each endpoint.5253Fetch all needed endpoints in parallel using curl or similar HTTP calls. All endpoints return JSON arrays or objects.5455## Step 3: Compute Results5657Use the formulas in [Formulas](references/formulas.md) to derive metrics from raw API data. Key computations include:5859- **Finance**: Revenue, COGS, gross margin, SG&A, allocations, EBITDA, EBITDA margin, ARPU, sales per labor headcount, growth rates60- **Compensation**: Pay type totals from rate book × roster weeks, quarter totals, overscale counts61- **Payroll**: Per-service type pay, category totals, premium calculations, conflict flag detection62- **Forecast**: Apply scenario growth rates to current-year compensation6364The [Formulas](references/formulas.md) reference covers every computation with worked examples. Do not improvise; these formulas are the single source of truth for all derived values.6566## Step 4: Format the Output6768Apply the output conventions in [Output Conventions](references/output-conventions.md):6970- Currency values rounded to 2 decimal places71- Percentage and ratio values rounded to 4 decimal places72- Lists ordered by ascending stable ID unless a rank field states otherwise73- All required top-level keys present in the exact structure from the answer template7475Write the result as a single JSON object. The task prompt will indicate where to save it (typically `answer.json` in the working directory).7677## Common Pitfalls7879- **Period mapping**: M1–M12 are FY2024, M13–M24 are FY2025. Do not confuse month numbers with period labels.80- **Revenue = product_revenue + service_revenue**: These are separate accounts; sum them.81- **COGS = direct_materials_cogs + direct_labor_cogs**: Same pattern.82- **SG&A = sales_sga + admin_sga + occupancy_sga**: Three accounts.83- **EBITDA = revenue - COGS - SG&A - allocations**: Not revenue - COGS - SG&A alone.84- **EBITDA margin = EBITDA / revenue**: Not divided by gross margin.85- **ARPU uses active_customers, not revenue_units**: `revenue / active_customers`.86- **Sales per labor headcount uses labor_headcount**: Not admin_headcount.87- **Growth rates**: `(new - old) / old`; ensure the denominator uses the older period.88- **Compensation**: Use roster quarter weeks (from the roster record), not a fixed 13-week default, for employees with partial-quarter schedules.89- **combined_overscale_includes_title**: When true, do not add a separate titled position premium for that employee; the overscale amount already includes it.90- **Forecast years**: Add one year of service for Year+1 and two years for Year+2 before assigning seniority bands.91- **Payroll rehearsal pay**: Hourly with a 3-hour minimum call duration.92- **Vacation**: 4% of base service pay plus premiums, only when `vacation_eligible` is true.93- **Payroll guarantee adjustment**: Applies only to regular players (not substitutes) when base service pay is below the weekly guarantee amount.94- **Substitute adjustment**: Deduct substitute service pay (audit + performance + doubles) from each regular musician's total. Add substitute pay as a separate line item for the substitute musician only.