Token导航 LogoToken导航TokenDH.com
图像处理需要联网github未标认证来源可访问许可证需确认审计通过

flux-analysis通量分析

Agent Skill

用于辅助图像生成、图片编辑、视觉素材处理或图像模型工作流。它适合让 Agent 根据文本生成图片、处理背景、整理视觉提示词或调用相关图像工具。使用时需要确认输入图片、版权来源、输出格式和模型限制;涉及人物、品牌、商品或公开展示素材时,应额外核对授权、真实性和内容合规边界。

总安装

423

周安装

18

GitHub Stars

55

下载量

148
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:flux-analysis(通量分析)
来源仓库:https://github.com/vm0-ai/vm0-skills
仓库路径:skills/flux-analysis
安装命令:
npx skills add https://github.com/vm0-ai/vm0-skills --skill flux-analysis
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/vm0-ai/vm0-skills --skill flux-analysis

简介

flux-analysis 用于辅助图像生成、图片编辑、视觉素材处理或图像模型工作流。

  • 适合让 Agent 根据文本生成图片、处理背景、整理视觉提示词或调用相关图像工具。
  • 使用时需要确认输入图片、版权来源、输出格式和模型限制。
  • 涉及人物、品牌、商品或公开展示素材时应额外核对授权、真实性和内容合规边界。
  • 适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

SKILL.md

Decomposition Techniques

Price x Volume Separation

The foundational split for any metric expressible as unit price multiplied by quantity.

Two-factor formulas:

Total Movement = Actual Result - Baseline (budget or prior period)

Quantity Component  = (Actual Units - Baseline Units) x Baseline Unit Price
Pricing Component   = (Actual Unit Price - Baseline Unit Price) x Actual Units

Check: Quantity Component + Pricing Component = Total Movement
       (when the interaction term is absorbed into one of the two factors)

Three-factor formulas (isolating composition shifts):

Quantity Component    = (Actual Units - Baseline Units) x Baseline Price x Baseline Mix Weights
Pricing Component     = (Actual Price - Baseline Price) x Baseline Units x Actual Mix Weights
Composition Component = Baseline Price x Baseline Units x (Actual Mix Weights - Baseline Mix Weights)

Worked example — top-line revenue:

  • Baseline plan: 10,000 units at $50/unit = $500,000
  • Actual outcome: 11,000 units at $48/unit = $528,000
  • Net movement: +$28,000 favorable

- Quantity uplift: +1,000 units x $50 = +$50,000 (favorable — higher volume) - Pricing drag: -$2 x 11,000 = -$22,000 (unfavorable — reduced average selling price)

Blended Rate / Composition Separation

Applicable when aggregated results blend multiple segments with distinct unit economics.

Formulas:

Rate Component = Sum across segments of [Actual Volume_i x (Actual Rate_i - Baseline Rate_i)]
Composition Component = Sum across segments of [Baseline Rate_i x (Actual Volume_i - Proportional Volume_i at Baseline Mix)]

Worked example — gross margin compression:

  • Segment X earns 60% margin; Segment Y earns 40% margin
  • Plan assumed a 50/50 split -> blended 50% margin
  • Actual split was 40/60 -> blended 48% margin
  • The 2-point margin decline is attributable to the shift toward the lower-margin segment

People-Cost Decomposition

Purpose-built for analyzing compensation and headcount-driven expense lines.

Total Compensation Movement = Actual Spend - Planned Spend

Break into:
1. Staffing level effect     = (Actual Headcount - Plan Headcount) x Plan Average Cost
2. Per-capita cost effect    = (Actual Average Cost - Plan Average Cost) x Plan Headcount
3. Composition effect        = Residual from shifts in seniority, department, or geography mix
4. Phasing effect            = Impact of hires arriving earlier or later than the plan assumed
5. Attrition benefit         = Savings from unplanned departures (partially offset by replacement and vacancy costs)

Functional Expense Decomposition

For operating cost categories where a price-times-volume model does not apply naturally.

Total OpEx Movement = Actual Operating Costs - Planned Operating Costs

Segment into:
1. Headcount-linked costs       (wages, benefits, payroll taxes, recruiting fees)
2. Activity-linked costs        (cloud hosting, payment processing fees, sales commissions, freight)
3. Discretionary programs       (travel, conferences, outside services, campaign spend)
4. Committed / fixed costs      (facility leases, insurance, enterprise software licenses)
5. Non-recurring charges        (severance, litigation, asset write-downs, project-specific outlays)
6. Timing / phasing differences (spend that shifted between periods relative to the plan)

Significance Thresholds & Prioritization

Calibrating Thresholds

Thresholds govern which movements warrant formal investigation and written explanation. Base them on:

  1. Overall materiality: Usually 1-5% of a primary benchmark (revenue, total assets, or net income)
  2. Relative line-item scale: Apply tighter percentage gates to larger balances
  3. Historical volatility: Allow wider bands for inherently variable accounts to filter noise
  4. Decision relevance: Would this size of movement influence a management decision or board discussion?

Suggested Threshold Grid

Comparison BasisSuggested Dollar GateSuggested Percentage GateTrigger Logic
Actual vs. annual planEntity-specific10%Whichever is breached first
Actual vs. same period last yearEntity-specific15%Whichever is breached first
Actual vs. latest forecastEntity-specific5%Whichever is breached first
Sequential month-over-monthEntity-specific20%Whichever is breached first

*Set the dollar gate at roughly 0.5-1% of revenue for income-statement lines.*

Triage Order When Multiple Items Exceed Thresholds

  1. Greatest absolute dollar impact — largest influence on the bottom line
  2. Greatest percentage deviation — may signal a process breakdown or data error
  3. Counter-trend movements — direction opposite to what history or forecasts predicted
  4. Newly emerged variances — items previously on track that have just diverged
  5. Compounding variances — gaps that have widened in each of the last several periods

Writing Effective Variance Narratives

Recommended Structure

[Line Item]: [Favorable / Unfavorable] movement of $[amount] ([X]%)
relative to [budget / prior period / forecast] for [reporting period]

Primary driver: [Brief label]
[Two to three sentences explaining the business cause, quantifying each
contributing factor where possible.]

Outlook: [One-time event / Likely to persist / Improving / Worsening]
Next step: [No action / Monitor / Deeper review / Adjust forecast]

Quality Criteria

Strong narratives consistently satisfy these tests:

  • Precise: Names concrete factors rather than restating the variance itself
  • Measured: Attaches dollar or percentage weight to each cited driver
  • Explanatory: Addresses why the movement occurred, not merely that it did
  • Prospective: States whether the movement is expected to continue, reverse, or evolve
  • Directive: Identifies any follow-up action or decision prompted by the finding
  • Compact: Two to four sentences — not padded filler

Pitfalls to Avoid

  • Restating the outcome as its own cause ("Revenue rose because revenue was higher")
  • Labeling a variance as "timing" without specifying what shifted and when normalization is expected
  • Calling something "one-time" without describing the actual event
  • Sweeping a material movement under "various small items" instead of decomposing further
  • Explaining only the dominant driver while ignoring meaningful offsets
  • Using vague qualifiers ("elevated," "slightly higher") without attached numbers

Bridge / Waterfall Presentation

Concept

A bridge (waterfall) chart traces the path from a starting value to an ending value through a sequence of additive and subtractive contributors. It is the visual companion to variance decomposition.

Data Architecture

Starting point:   [Baseline figure — plan, prior period, or forecast]
Contributors:     [Ordered list of signed driver amounts]
Ending point:     [Actual figure]

Integrity check:  Starting point + Sum(all contributors) = Ending point

Text-Format Bridge (When No Charting Tool Is Available)

BRIDGE: Operating Expenses — Q4 Actual vs. Q4 Plan

Q4 Planned OpEx                                       $8,000K
  |
  |--[+] Incremental headcount above plan              +$500K
  |--[+] Unplanned outside counsel fees                +$200K
  |--[-] Open-role savings (delayed hiring)            -$350K
  |--[-] Travel spend below budget                     -$180K
  |--[+] Cloud infrastructure overrun                  +$130K
  |--[-] Marketing program deferrals                   -$100K
  |
Q4 Actual OpEx                                        $8,200K

Net Movement: +$200K (+2.5% unfavorable)

Companion Reconciliation Table

DriverAmountShare of Total MovementRunning Total
Incremental headcount+$500K250%+$500K
Outside counsel+$200K100%+$700K
Open-role savings-$350K-175%+$350K
Travel underspend-$180K-90%+$170K
Cloud overrun+$130K65%+$300K
Marketing deferrals-$100K-50%+$200K
Net movement+$200K100%

*Individual shares can exceed 100% when favorable and unfavorable drivers offset each other.*

Presentation Guidelines

  1. Sequence drivers from most favorable to most unfavorable (or in a logical business narrative order)
  2. Cap the driver count at 5-8; roll smaller items into an "All other" bucket
  3. Verify arithmetic: opening value plus all drivers equals closing value
  4. Use color to distinguish direction — green for favorable, red for unfavorable — in graphical renderings
  5. Annotate each segment with both the dollar amount and a short label
  6. Include a summary segment showing the net total movement

Multi-Scenario Comparisons

Three-Column Layout

Line ItemAnnual PlanLatest ForecastActualPlan Var ($)Plan Var (%)Forecast Var ($)Forecast Var (%)
Revenue$X$X$X$XX%$XX%
Direct costs$X$X$X$XX%$XX%
Gross profit$X$X$X$XX%$XX%

Choosing the Right Baseline

  • Actual vs. annual plan: Governance and incentive evaluation; the plan is fixed at the start of the fiscal year
  • Actual vs. rolling forecast: Operational steering and early-warning detection; the forecast is refreshed monthly or quarterly
  • Forecast vs. plan: Gauges how management expectations have shifted since planning; highlights planning-accuracy gaps
  • Actual vs. prior period (sequential): Reveals trend direction; especially useful for new ventures or post-acquisition integration where a plan may not yet exist
  • Actual vs. prior year (year-over-year): Growth assessment adjusted for seasonality

Tracking Forecast Precision

Measure forecast quality over time to tighten future planning:

Period Accuracy = 1 - |Actual - Forecast| / |Actual|

MAPE (Mean Absolute Percentage Error) = Mean of |Actual - Forecast| / |Actual| across all periods
MonthForecastActualDeviationAccuracy
Jan$X$X$X (X%)XX%
Feb$X$X$X (X%)XX%
...............
Full YearMAPEXX%

Reading Variance Trends Across Time

  • Persistently favorable: Plans may be overly conservative (potential sandbagging)
  • Persistently unfavorable: Targets may be unrealistic or execution is lagging
  • Widening unfavorable gap: Performance is deteriorating or external conditions are shifting
  • Narrowing gap: Forecast accuracy is improving through the year (a healthy signal)
  • Erratic swings: Business is inherently unpredictable or the forecasting methodology needs refinement

适合场景

01

文本生成图片

02

图片风格化

03

产品图和创意图

04

需要 FLUX 模型时

能力概览

能力 1

调用 FLUX 图像模型

能力 2

支持文本生图和图像改写

能力 3

覆盖 LoRA 或风格适配

能力 4

适合创意视觉生成

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

36.45%
按下载量换算54

Claude

31.21%
按下载量换算46

Cursor

18.91%
按下载量换算28

Gemini CLI

9.16%
按下载量换算14

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills