MERINO EXPERIMENTS Lab 01 · Entropy collapse
Defensive research Offline ₿ Testnet
MERINO / LAB 01 Weak entropy under observation

The output looks
128-bit. The generator isn’t.

Generate a deliberately vulnerable Bitcoin testnet wallet, expose realistic operational constraints, and watch the reachable state space collapse into a searchable region.

Core lesson Hashing can hide bias. It cannot manufacture entropy.
01 / SPECIMEN

Vulnerable wallet ceremony

Deliberately unsafe

No specimen generated

The laboratory creates a fixed demonstration fixture. It never accepts wallet data.

! Prototype fixture only. Never enter a real seed, address, xpub, or passphrase.

02 / ENTROPY FUNNEL

What does the attacker learn?

Current search 41.3 bits
Apparent BIP39 entropy2128
Visually uniform
Reachable generator states2.72T
41.3 bits
After selected constraints2.72T
2,717,908,992,000 candidates
Constraint evidence Select every justified observation
⌁
Search region is too large Apply evidence until the executable region is below 65,536 candidates.
2.72T
03 / RECOVERY INSTRUMENT

Enumerate states. Derive addresses. Test the oracle.

Every candidate follows the complete presentation pipeline: deterministic state → BIP39 → BIP32 → BIP84 testnet address.

Search engine
Tested00.0%
Candidate rate—Measured workers
Elapsed00:00.0Wall clock
Matches0Exact address
Candidate topology Generation minute × UID suffix
Unseen Tested Match
Fake UID suffix →
Generation time →
CONTROL / SECURE CONSTRUCTION

Change the source,
not the appearance.

Independent entropy128+ bits

Use an operating-system CSPRNG or sufficient private, fair dice input. The same BIP39 output format now represents a candidate space the demonstration cannot enumerate.