Agent Skills
› cosmicstack-labs/mercury-agent-skills
› budgeting-forecasting
budgeting-forecasting
GitHub提供零基预算、滚动预测、差异分析和情景规划等财务管理方法，指导用户创建和管理有效预算及财务预测，支持成本优化与决策。
触发场景
制定年度或月度预算
进行财务差异分析
构建滚动预测模型
制定情景规划方案
安装
npx skills add cosmicstack-labs/mercury-agent-skills --skill budgeting-forecasting -g -y
SKILL.md
Frontmatter
{
"name": "budgeting-forecasting",
"metadata": {
"tags": [
"budgeting",
"forecasting",
"finance",
"planning",
"variance"
],
"author": "cosmicstack-labs",
"version": "1.0.0",
"category": "finance-legal"
},
"description": "Zero-based budgeting, rolling forecasts, variance analysis, scenario planning, and budget management"
}
Budgeting & Forecasting
Create and manage effective budgets and financial forecasts.
Budgeting Approaches
| Approach | Method | Best For |
|---|---|---|
| Incremental | Last year + X% | Stable organizations |
| Zero-based | Build from scratch each period | Cost optimization |
| Activity-based | Cost per unit of activity | Manufacturing, services |
| Rolling | Continuous 12-month outlook | Fast-changing environments |
Zero-Based Budgeting Steps
- Identify all activities and their costs
- Evaluate each: does it create value?
- Rank by priority/cost ratio
- Fund top priorities within budget constraints
- Cut or optimize low-value activities
Forecasting
Rolling Forecast Model
# Structure
- Historical data (12-24 months)
- Current pipeline and commitments
- Leading indicators (traffic, leads, bookings)
- Economic assumptions
- Scenario overlays (best/base/worst)
Variance Analysis
| Variance | Cause | Action |
|---|---|---|
| Revenue ↑ | Volume or price increase | Sustain driver or reinvest |
| Revenue ↓ | Volume or price drop | Investigate root cause |
| Cost ↑ | Variable or fixed increase | Optimize or renegotiate |
| Cost ↓ | Efficiency or lower demand | Validate sustainability |
Scenario Planning
- Base case: Most likely outcome
- Upside: +20% revenue, what would we invest in?
- Downside: -20% revenue, what would we cut?
- Each scenario has specific triggers and action plans
- Review scenarios monthly against actuals
版本历史
- 38e2523 当前 2026-07-05 19:39
