Nice idea—scavenged parts, self‑repair, and a ledger on a swarm. First off, keep the ledger small. A simple Merkle tree on each node with a few hash chains works; no need for a full blockchain, just enough to catch tampering. Use a gossip protocol for fault reports: each drone pushes its log entry to a handful of neighbors, and the gossip rounds ensure the whole swarm eventually sees it. Make the communication mesh self‑healing too. If a node drops out, the others automatically re‑route their messages through the next hop. Use a lightweight discovery service—maybe a simple multicast hello, then keep a local table of active peers. For self‑repair, expose a micro‑service on each drone that can fetch spare parts from the swarm. When a fault is logged, the service can pull the needed component from the nearest node with a surplus. Keep the inventory list on the same ledger so everyone knows who can supply what. Finally, don’t forget to throttle the gossip to avoid flooding the network. A small TTL and a random back‑off on re‑broadcasts will keep the swarm stable. Good luck, and keep an eye on the edge cases—those rogue nodes love to break things.