A tool economy, or tool ecology, is itself a highly modular machine with hardware, that is, an infrastructure, and software, that is, processes that run on the infrastructure. The hardware/software aspect is not accidental; if one looks at Babbage's plans for the Analytical Engine, one quickly realizes that Babbage was modeling computation on production: the Analytical Engine is a little tool ecology or manufacturing system of its own, for producing a particular kind of product (data structures) from a particular kind of supply (energy). Any tool economy is itself a sort of Engine, analytical or not.
I was interested to come across the Global Village Construction Set, which is an attempt to pull together into one package a little sustainable manufacturing system of its own -- Industrial Revolution in a box, so to speak. It's based on the idea of product ecologies: you can use (say) solar collectors to power steam engines to generate electricity to run machining tools that can make solar collectors, steam engines, electrical generators, and machining tools, thus creating a potentially sustainable cycle. If you can link several different possible cycles, you have a network and a highly flexible system. The GVCS is planned to make use of 50 different machine components:
CEB Press, Cement Mixer, Saw Mill, Bulldozer, Backhoe
Tractor, Seeder, Hay Rake, Well Drilling Rig
Microtractor, Soil Pulverizer, Spader, Hay Cutter, Trencher
Bakery Oven, Dairy Milker, Microcombine, Baler
Multimachine, Ironworker, Laser Cutter, Welder, Plasma Cutter, Induction Furnace
Torch Table, Metal Roller, Rod and Wire Mill, Press Forge, Universal Rotor, Drill Press
3D Printer, 3D Scanner, Circuit Mill, Industrial Robot, Chipper Hammermill
Power Cube, Gasifier Burner, Solar Concentrator, Electric Motor Generator, Hydraulic Motor, Nickel-Iron Battery
Steam Engine, Heat Exchanger, Wind Turbine, Pelletizer, Universal Power Supply
Aluminum Extractor, Bioplastic Extruder
Most of it is still in the planning and development stages. It's one thing, of course, to have these things, which we obviously do, and another to have them under the right conditions: the project requires cheap, replicable versions that are capable of working with the other parts in the system to allow actual replication. Not an easy thing to do. But some of the basic ideas are quite interesting.
To some degree, though, this is due to the ambition of the project. One could imagine less ambitious versions, intended not as tool micro-economies but as upgrade systems for tool economies already in place. Further, for a lot of things one wouldn't need state-of-the-art solar collectors and the like; technology far short of the state of the art can do less, but it also generally requires less to replicate it and sustain its tool economy. And one could also focus on specific subdomains of the tool economy, like agriculture, irrigation, or basic tool-making. In a sense this is the kind of thing that the Africa Windmill Project is trying to put together.