Mission & System Cost Estimation Skill

Read CONVENTIONS.md at the repo root before proceeding.

This skill generates parametric cost estimates for space missions — from early ROM for proposals and investor pitches through detailed cost breakdowns for Phase A/B design reviews.

Before You Begin

Ask the user (if not already known):

What is being costed? — Full mission (spacecraft + launch + ground + ops), spacecraft bus only, single subsystem, or a component?
What cost class? — NewSpace/commercial (lean teams, COTS parts, agile development) or Traditional/institutional (full documentation, Class B/C, government oversight)?
What is the mission type? — Earth observation, comms, science, lunar, interplanetary, constellation, crewed? (Different CERs apply)
What cost year (base year)? — All costs must be stated in a specific fiscal year (e.g., FY2026$). Apply NASA New Start Inflation Index or GDP deflator for conversion.
What design phase? — Phase A: ROM ±50%; Phase B: ±30%; Phase C: ±15%.
Is there a target cost or cost cap? — Design-to-cost constraints change the analysis approach.

Applicable Phases

Primary: Phase A (ROM for proposals, mission feasibility), Phase B (refined estimates for confirmation review)
Supporting: Phase C (cost-at-completion tracking), Phase D (actuals reconciliation)

Ownership Boundary

| Responsibility | Owner | |:---|:---| | Subsystem mass and complexity inputs | Domain analysis skills (structural-assessment, power-assessment, etc.) | | System-level cost estimate, cost model selection, wraps, and risk | This skill | | Trade study cost criterion scoring | trade-study-manager (uses cost estimates from this skill) | | Schedule and program timeline | schedule-risk-assessment (when available) |

Cost Estimation Methodology

1. Select the Cost Model

Choose the appropriate model based on mission type and available data:

| Model | Best For | Key Input | Notes | |:---|:---|:---|:---| | USCM8/9 (Unmanned Spacecraft Cost Model) | Traditional government S/C | Dry mass by subsystem | NASA/Air Force heritage, well-calibrated | | SSCM (Small Satellite Cost Model) | SmallSats < 500 kg | Total dry mass, mission type | Aerospace Corp, better for small missions | | NICM (NASA Instrument Cost Model) | Science instruments/payloads | Instrument mass, type, aperture | Separate from bus cost | | PCEC (Project Cost Estimating Capability) | NASA missions (full lifecycle) | WBS-level inputs | NASA's primary tool | | Analogy | When a heritage mission exists | Historical actual cost + adjustments | Best when strong analog exists | | NewSpace Parametric | Commercial/startup missions | Mass, TRL, team size, reuse factors | Calibrated to commercial actuals |

Default approach: If the user doesn't specify, use mass-based parametric CERs (USCM-class) with NewSpace adjustment factors where applicable.

2. Spacecraft Bus Cost (Hardware)

Estimate each subsystem using mass-based CERs. Generic form:

$$C_{subsystem} = a \cdot M^b \cdot F_{complexity}$$

Where:

$M$ = subsystem mass (kg)
$a$, $b$ = model coefficients (subsystem-specific)
$F_{complexity}$ = complexity adjustment (1.0 = average, 0.7 = COTS/high-heritage, 1.5 = novel/custom)

Reference CER Coefficients (USCM-class, FY2020$K)

| Subsystem | a | b | Typical Mass Range | |:---|:---|:---|:---| | Structure & Mechanisms | 157 | 0.83 | 10-500 kg | | Thermal Control | 394 | 0.635 | 2-50 kg | | EPS (Power) | 62.7 | 1.00 | 5-200 kg | | TT&C (Communications) | 545 | 0.761 | 2-50 kg | | ADCS (GNC) | 464 | 0.867 | 3-100 kg | | Propulsion | 18.4 | 0.846 | 5-500 kg | | C&DH (Flight Computer) | 545 | 0.761 | 2-30 kg |

⚠️ These are reference values only. Actual CER coefficients vary by source database and should be stated with their provenance. Ask the user if they have project-specific CERs.

NewSpace Adjustment Factors

For commercial/startup missions, apply reduction factors to traditional CERs:

| Factor | Typical Range | Rationale | |:---|:---|:---| | COTS Hardware | 0.3 - 0.6 | Commercial off-the-shelf vs. custom flight hardware | | Lean Team | 0.5 - 0.7 | Small agile team vs. large institutional workforce | | Reduced Documentation | 0.7 - 0.85 | Streamlined reviews vs. full NASA/ECSS doc suite | | Design Reuse / Heritage | 0.4 - 0.7 | Block-buy, repeat build, product line | | Composite NewSpace Factor | 0.15 - 0.40 | Combined effect (multiply individual factors) |

3. Payload / Instrument Cost

Use NICM or analogy:

Optical instruments: $C = f(\text{aperture diameter}, \text{mass}, \text{detector type})$
RF instruments (SAR, radiometers): $C = f(\text{antenna area}, \text{power}, \text{bandwidth})$
In-situ (spectrometers, drills): Best estimated by analogy to heritage instruments.

Rule of thumb: Payload cost is typically 20-40% of spacecraft bus cost for observatory missions, but can exceed bus cost for flagship science instruments.

4. Wrap Factors (Non-Hardware Costs)

Hardware cost alone dramatically underestimates total mission cost. Apply wrap factors:

| WBS Element | Typical % of Hardware Cost | Notes | |:---|:---|:---| | Program Management (PM) | 8-15% | Lower for small missions, higher for institutional | | Systems Engineering (SE) | 10-18% | Includes budgets, interfaces, trade studies | | Mission Assurance (MA) | 3-8% | Quality, reliability, parts screening | | Integration & Test (I&T) | 10-20% | AIT campaign, environmental testing, GSE | | Ground Segment | 15-30% of total | MOC, ground stations, data processing | | Launch Services | Market price | SpaceX rideshare: ~$5.5K/kg; Dedicated small LV: $10-30M; Medium/Heavy: $60-150M | | Operations (per year) | 5-15% of S/C cost per year | Staff, ground network fees, data processing |

5. Mission Lifecycle Cost

Assemble the total:

$$C_{total} = C_{payload} + C_{bus} + C_{PM/SE/MA} + C_{I&T} + C_{launch} + C_{ground} + C_{ops} + C_{reserve}$$

Cost Reserve (Unallocated Future Expenses — UFE):

Phase A: 30-50% of estimated cost
Phase B: 20-30%
Phase C/D: 10-15%
NASA policy (NPR 7120.5): minimum 25% UFE at KDP-C for Class B missions.

6. Cost-Risk Analysis

Go beyond point estimates:

Confidence Levels

50th percentile (P50): Equal chance of overrun or underrun. Typical for proposals.
70th percentile (P70): Common for budgeting. NASA often uses P70 for Class B/C.
80th percentile (P80): Conservative. Used for cost caps and flagship missions.

S-Curve Generation

If uncertainty distributions are available:

Assign triangular or log-normal distributions to each subsystem cost (min, most likely, max).
Run Monte Carlo simulation (1000+ draws).
Plot cumulative probability vs. total cost = S-curve.
Read off P50, P70, P80 values.

Simplified approach (Phase A): Apply cost growth factors from historical data:

Average NASA mission cost growth: 1.4× from Phase A to actuals
Average commercial: 1.2× (smaller scope, fewer changes)

7. Design-to-Cost (DTC)

When the user specifies a cost cap:

Work backward from the cap to allocate subsystem budgets.
Identify which subsystems are cost-driving (typically payload, structure, and ADCS).
Flag if the cost cap is unrealistic given the mission requirements.
Recommend trades: reduce performance, increase COTS usage, descope payload, or change orbit.

Common Reference Costs (FY2025$, approximate)

For quick sanity checks and Phase A estimates:

| Mission Class | Typical Total Cost | Examples | |:---|:---|:---| | 3U CubeSat (university) | $0.5-2M | Educational, tech demo | | 6-12U CubeSat (commercial) | $2-10M | Planet Dove, Spire | | Microsatellite (50-150 kg) | $10-50M | ICEYE SAR, BlackSky | | Small satellite (150-500 kg) | $30-150M | RCM, WorldView Legion | | Medium satellite (500-2000 kg) | $150-500M | Sentinel, GOES-R instruments | | Large science mission | $500M-2B | JWST ($10B is an outlier) | | Flagship mission | $2-10B | Mars Sample Return, Europa Clipper | | Lunar lander (commercial) | $50-300M | CLPS landers (Astrobotic, Intuitive Machines) | | Constellation (per sat, at scale) | $0.5-5M | Starlink, OneWeb (unit cost at volume) |

⚠️ These are order-of-magnitude reference points. Actual costs vary enormously based on complexity, heritage, and programmatic approach. Always state the basis of estimate.

Output Format

1. Cost Estimate Report (`cost_estimate.md`)

A Markdown document containing:

Basis of Estimate (BoE): Model used, key assumptions, cost year, heritage reference
Cost Breakdown Table: By WBS element (hardware, PM, SE, I&T, launch, ground, ops)
Subsystem Cost Table: By spacecraft subsystem
Confidence Level: Point estimate and stated confidence (P50/P70/P80)
Cost Drivers: Top 3 cost-driving items
Cost-Risk Summary: Key uncertainties and their cost impact
🟢 / 🟡 / 🔴 Status: Green if within cap/target, yellow if within reserve, red if exceeds cap

2. Cost Comparison Table (`cost_comparison.csv`)

If multiple options are being compared (for trade-study-manager):

Side-by-side cost breakdown for each option
Lifecycle cost comparison (development + operations over mission life)

Verification & Cross-Checks

Mass-cost correlation: Verify that $/kg is within reasonable bounds for the mission class (typically $50K-500K/kg for traditional, $10K-100K/kg for NewSpace).
Historical analog: If possible, compare to a known mission of similar scope.
Subsystem proportions: Check that no single subsystem exceeds 40% of bus cost unless justified (e.g., a very large solar array).

Interface

Reads from: /requirements/, /analysis/systems-engineering-assessment/ (mass budget), /analysis/structural-assessment/ (mass properties), all domain skill outputs (for subsystem mass inputs)
Writes to: /analysis/cost-estimation/
Consumed by: trade-study-manager (cost as a Figure of Merit), systems-engineering-assessment (mission-level summary), project management / proposal teams