🚀 rocket-tools

Aerospace engineering intelligence for AI agents.

What is rocket-tools?

A dual-interface engineering computation engine:

• MCP Server — 11+ aerospace engineering tools callable by AI agents via the Model Context Protocol
• Skills Library — Human-readable .md references with formulas, worked examples, and pitfalls

Built for speed, precision, and composability.

Features

Domain	Tools	Description
Structural	beam_analysis	Bending, deflection, shear, Euler buckling — rectangle & circle sections
Materials	material_lookup	5 aerospace alloys with full thermal/mechanical properties
Atmosphere	isa_atmosphere	ISA 0–25,000 m with pre-computed cache + linear interpolation
Aerodynamics	reynolds_number, mach_number, dynamic_pressure, lift_coefficient, drag_coefficient, skin_friction_coefficient, aero_analysis	Flow regime, compressibility, forces
Utilities	unit_convert	NIST-traceable conversions (SI ↔ imperial, temperature)

Performance

Operation	Latency
ISA lookup (cached)	~54 ns
Material lookup	~342 ns
Unit conversion	~205–546 ns
Reynolds number	~1.4 μs
Beam analysis (Numba JIT)	~3.0 μs
Full aerodynamic analysis	~6.6 μs

All under the 1 ms per-tool target.

Quick Start

Install

pip install -e ".[dev]"

Run Tests

pytest -v                    # 34 tests
pytest --benchmark-only -v   # 18 benchmarks

Use as Python Library

from rocket_tools.aerodynamics import aero_analysis
from rocket_tools.structural import beam_analysis

# Comprehensive aerodynamic characterization
result = aero_analysis(
    velocity=250.0,
    altitude_m=5000.0,
    characteristic_length=20.0,
    reference_area=40.0,
    lift=50000.0,
    drag=5000.0,
)
print(f"Re = {result['reynolds_number']:.1e}")
print(f"Mach = {result['mach_number']:.3f} ({result['mach_regime']})")
print(f"L/D = {result['lift_to_drag_ratio']:.1f}")

# Beam design check
beam = beam_analysis(
    load=100.0,
    length=1.0,
    youngs_modulus=68.9e9,  # 6061-T6
    cross_section={"type": "rectangle", "width": 0.05, "height": 0.01},
)
print(f"Deflection: {beam['max_deflection_m']*1000:.3f} mm")
print(f"Bending stress: {beam['bending_stress_pa']/1e6:.2f} MPa")

Use as MCP Server

rocket-tools  # Starts FastMCP server

Then connect your MCP client (Claude, Cursor, etc.) to the running server. Available tools:

• unit_convert — Convert engineering units
• material_lookup — Look up material properties
• isa_atmosphere — ISA atmosphere at altitude
• beam_analysis — Structural beam analysis
• reynolds_number — Reynolds number calculation
• mach_number — Mach number at altitude
• dynamic_pressure — Dynamic pressure
• lift_coefficient — Lift coefficient
• drag_coefficient — Drag coefficient
• skin_friction_coefficient — Skin friction (Blasius)
• aero_analysis — Comprehensive aerodynamic analysis

Skills Library

Human-readable engineering references in skills/:

• skills/structural-analysis.md — Beam theory, Euler buckling, section properties
• skills/aerodynamics.md — Reynolds, Mach, dynamic pressure, lift/drag

Each skill includes:

• LaTeX formulas
• MCP tool cross-references
• Worked Python examples
• Common pitfalls

Architecture

rocket_tools/
├── utils/          # Units, validation, caching
├── materials/      # Material database + ISA atmosphere
├── structural/     # Beam mechanics (Numba JIT)
├── aerodynamics/   # Re, Mach, q, CL, CD, Cf (Numba JIT)
├── server.py       # FastMCP tool definitions
└── rust_kernels/   # Rust PyO3 extension (scaffolded)

Numba JIT accelerates all hot paths. Rust kernels are scaffolded for future PyO3 integration when network/toolchain issues resolve.

Project Roadmap

Phase	Status	Features
Phase 1	✅ Complete	Core tools (structural, materials, aerodynamics), tests, benchmarks, skills
Phase 2	🔄 Next	Natural language router, composable workflows, uncertainty propagation, contextual memory
Phase 3	📋 Planned	Visual intelligence (plots/diagrams), design optimization, standards compliance
Phase 4	📋 Planned	Knowledge graph, FMEA, multi-agent sessions, plugin architecture

Contributing

1. Fork the repo
2. Create a feature branch
3. Write tests first (TDD)
4. Ensure pytest -v passes
5. Commit and push

License

Apache-2.0 — See LICENSE

---

Built by Human Engine for the agentic era.