A confirmed Workshop round-trip
Fund, split, and consolidate on public Signet: 7,964 sats returned to the source wallet, with 1,298 sats in total fees.
Bitcoin Inquisition 29.2.0 and LND 0.20.1-beta / default Signet
Reproducible experiments
Transaction reports
Worked examples, serialized artifacts, and recorded node outcomes. Each report names its actual test environment and the limits of its result.
Node-tested experiments
When a fee rescue works
Five Core scenarios cover an insufficient replacement fee, an accepted replacement, package accounting, and the confirmed-parent boundary.
Core 29.0.0 / isolated regtest
What authorizes a spend?
Eight recorded outcomes compare early and eligible timelocks, missing signatures, a wrong preimage, and a delayed refund.
Core 29.0.0 / isolated regtest
The signature that does not replay
A six-case matrix tests ordinary and unified signatures across a shared-history fork, with block acceptance recorded separately.
Core 29.0.0 and Knots 29.4.1 / isolated regtest
One input, three recipients, and the change that remains
Account for every satoshi in a native SegWit split, including the fee and a separate change output.
Consolidation has an input cost
Measure a three-input, one-output consolidation and see why gathering coins is not automatically a fee saving.
A PSBT can parse while its fee remains unknown
Remove previous-output metadata, inspect the missing information, and compare it with the complete unsigned PSBT.
Unsigned, signed, finalized: three different states
Inspect what is present at each PSBT stage without mistaking a signature field for a validated transaction.
What counts as evidence here?
Each report exposes its transaction data and construction assumptions. An offline serialization test establishes the bytes and the local calculation. It does not establish network acceptance, ownership of an input, or confirmation. Those outcomes require separately recorded node or chain evidence.
Download the fixture collection · Source and reproduction scripts