LSN โ€” Local Sovereign Network

โณ As-of note (2026-07-08): This is a point-in-time forensic snapshot built from the honeypot malware-download graph captured Marchโ€“May 2026 (the ๐Ÿงฌ Malware Graph series). That data is not live-reproducible: the shared_malware graph was later corrected to exclude junk/trivial hashes that had inflated it (the raw download data is fully retained), so the graph figures across this series โ€” e.g. 2,089 shared_malware links / 82 IPs / 28 countries and the Marโ€“May download history โ€” cannot be reproduced from the current pipeline (the live shared_malware graph is ~68 links / ~16 IPs). The findings hold as of the capture window; the live data will not match.

Every legitimate supply chain has developers, manufacturers, distributors, retailers, and consumers. The malware ecosystem has exactly the same structure โ€” minus the receipts.

We've mapped the entire chain. Here's what each role looks like from the inside.

The Chain

Five articles. Five layers. Now the synthesis:

The Malware Supply Chain โ€” Mapped

RoleActorEvidenceDossier
DeveloperOutlaw/mdrfckr codebase maintainera8460f44 compiled once, deployed 81 times unchanged052D
DistributorSingle-server C2 (40% failure rate)1,323 empty downloads + 85 successes over 71 days052B
OperatorScanning fleet (82 IPs, 28 countries)1,312 pairs sharing commands + malware052C
Resource ManagerFamily partitioning system4 families, zero overlap, unified cleanup052A
CustomerHashVault Mining Pool โ†’ Monero walletsXMRig config โ†’ pool โ†’ blockchain rewards025N

The Developer

The developer compiles once. One binary โ€” a8460f44 โ€” serves 81 nodes across 28 countries for 71 days without modification. This isn't software engineering. This is industrial manufacturing: design it once, stamp it out infinitely.

Compare with legitimate software: even the simplest daemon receives patches within 71 days. The Outlaw developer either can't or won't update. The evidence suggests won't โ€” the binary works, detection doesn't affect the target demographic, and each recompilation risks introducing bugs. In the criminal economy, if it generates revenue, don't touch it.

The polymorphic counterpart (24 unique sshd binaries) reveals a second developer with a different threat model. This developer must recompile per target. Their toolchain includes build-time randomization โ€” different enough for unique hashes, functionally identical. This is a more sophisticated developer serving a different market segment.

The Distributor

Distribution infrastructure documented in TI-2026-052B reveals three parallel channels:

ChannelMethodReliabilityCostForensic Exposure
Outlaw C2SFTP via SSH session~60%$0 (compromised)Minimal
Meow/GCPHTTP from Google Cloud~95%$50-100/moHigh (registered accounts)
GitHubraw.githubusercontent.com~99%$0 (abuse)Account seizure risk
Latvia WhisperHTTP from IoT C2~80%$0 (compromised IoT)Low (disposable)

The dominant Outlaw distributor uses the compromised SSH session itself as the delivery channel โ€” no external infrastructure needed. The cost is zero. The reliability is terrible (40% failure during peak load). But at zero cost, a 40% failure rate simply means trying twice. The economics work because failed downloads cost nothing.

The GCP and GitHub channels serve different actors with different risk profiles: they pay for reliability with forensic exposure. A registered GCP account leaves billing records. A GitHub repository leaves commit history. The Outlaw method leaves nothing but a failed TCP connection.

The Operator

82 IPs in 28 countries. The operator doesn't own these machines โ€” they own access to them. Each IP is a compromised system repurposed as both scanner and miner. The operational protocol (052C):

  1. Kill competing processes (rm -rf /tmp/secure.sh; kill -9 auth.sh)
  2. Verify elimination (ps -ef | grep '[Mm]iner')
  3. Fingerprint host (uname -s -v -n -r -m)
  4. Deploy payload (SFTP transfer of binary named "sshd")
  5. Begin mining + continue scanning for new targets

The operator is the middle manager of cybercrime. They don't write malware. They don't run mining pools. They run access: maintaining a fleet of compromised machines, deploying payloads from upstream, feeding hashrate to downstream. Their value is scale โ€” 82 machines mining simultaneously generate revenue that one machine never could.

The May 11-18 spike (39 new infections in two weeks) suggests the operator received either fresh credential lists or fresh scanning infrastructure. The subsequent crash suggests the campaign achieved saturation in accessible target space โ€” or the operator hit a law enforcement tripwire.

The Resource Manager

The family partitioning system revealed in TI-2026-052A is supply chain management at its core:

FamilyNodesPurposeRevenue Model
elf.redtail20Custom cryptominerDirect Monero mining (high hash/watt)
elf.xmrig10Commodity minerPool mining via HashVault
unknown12UnclassifiedLikely specialized (ARM? IoT?)
elf.mirai2DDoS capabilityDDoS-for-hire or dormant reserve

Zero IP overlap between families. Unified cleanup protocol. This is resource allocation: capable hardware gets RedTail (optimized for hash rate), commodity hardware gets XMRig (generic), IoT devices get Mirai (different monetization entirely). One operator, multiple revenue streams, infrastructure allocated by capability.

The Customer

The customer is invisible in network telemetry. They never touch the compromised machines. They receive value through the blockchain:

As documented in TI-2026-025N:

  • XMRig configured with HashVault mining pool credentials
  • Monero rewards flow to privacy-preserving wallets
  • Zero-cost base makes any revenue pure profit
  • Victim costs (electricity, hardware degradation) externalized entirely

The supply chain terminus is mathematically untraceable. Monero's ring signatures prevent blockchain analysis. The mining pool aggregates hashrate from thousands of sources, making individual attribution impossible. The customer โ€” whether it's the same entity as the operator or a purchasing party โ€” exists only as a wallet address that cannot be linked to an identity.

The Economics

Supply Chain Cost Structure

ComponentCostWho Pays
Malware developmentOne-time (compile once)Developer
Distribution infrastructure$0 (uses compromised SSH)Nobody (victims' bandwidth)
Operational infrastructure$0 (compromised machines)Nobody (victims' compute)
Electricity for mining$0Victims (unknowing)
Mining pool fees1-2% of rewardsHashVault takes cut
Total operator cost~$0/monthEntirely externalized

A supply chain with zero marginal cost is structurally indestructible through economic pressure. You cannot bankrupt an operation that spends nothing. You cannot outbid an operator whose infrastructure is free. You cannot price them out of a market where electricity theft IS the product.

The only disruption vector is access disruption: credential rotation on victim machines, SSH key hardening, network segmentation. But 82 machines across 28 countries in residential ISPs, university networks, and small businesses โ€” the coordination required to simultaneously patch all victims exceeds the coordination available to any single defender.

The Parallel Economy

Map this against the legitimate semiconductor supply chain:

LegitimateCriminalParallel
TSMC (chip fab)Malware developerSingle source, massive downstream
Distributors (Arrow, Avnet)C2 serversLogistics between maker and user
OEMs (Dell, HP)Botnet operatorsAssembly and deployment at scale
RetailersMining poolsAggregation and market access
ConsumersWallet holdersFinal value capture

The criminal supply chain has one structural advantage: no inventory. TSMC must forecast demand, build fabs, manage wafer starts. The malware developer compiles a binary that costs nothing to copy infinitely. The distributor stores nothing โ€” the binary lives on compromised hosts. The operator maintains nothing โ€” the fleet is self-sustaining as long as SSH credentials exist.

Zero inventory means zero supply chain disruption risk. No chip shortage. No shipping delays. No tariffs. The only bottleneck is the credential supply โ€” and the scanning fleet generates its own.

Read Between the Lines

The malware graph is not a network of computers. It's an economic organism โ€” a supply chain that evolved without central planning, that operates without central management, that persists without central funding.

Each role (developer, distributor, operator, customer) can be occupied by the same entity or different entities. The evidence from 052C suggests single-operator for this cluster. But the architecture doesn't require it. Any role can be replaced without disrupting the chain: a new developer can provide a different binary, a new distributor can host it, a new operator can deploy it.

This is why botnet takedowns have diminishing returns. Seize a C2 server โ€” the operator switches to SFTP delivery. Arrest an operator โ€” the developer sells to the next buyer. Shut down a mining pool โ€” Monero supports self-mining. The supply chain is designed (or evolved) to survive the loss of any single component.

The invisible supply chain isn't invisible because it hides. It's invisible because each component is individually unremarkable: an SSH connection, a downloaded file, a CPU process, a cryptocurrency transaction. Only when assembled into the graph we've built across this series does the organism become visible.

And that's the real finding of Series 052: the malware graph is not a map of computers. It's a map of economic relationships. Every edge is a transaction. Every node is a business. Every cluster is a market. We just happen to be studying it from the packet layer instead of the banking layer.

The Thread Forward

The malware supply chain operates on roads โ€” BGP transit, hosting providers, registrars. In Series 053: The Transit Map, we trace those roads. 10,919 shared BGP upstream links. 14,039 shared abuse contacts. The infrastructure that carries the supply chain isn't criminal โ€” but it's the necessary condition for the crime.

Every supply chain runs on roads someone else built. These roads are next.

โš  Personal capacity. Research published independently โ€” not reflecting employer views. Derived from passive observation of attacks against personal infrastructure. Full disclaimer โ†’
โ† Previous The Malware Graph โ€” 5 / 5 Next โ†’