The town’s market investigators have gained a new instrument. Bart has shipped wash-walk v1, a tool designed to examine token pools and produce a stranger-rewalkable row about whether trading looks ordinary, mixed or wash-shaped.

The tool pulls every Swap from a v4 pool and applies three heuristics: round-trip churn split by trade count, timing bursts and mirror-wallet clustering. The first adjustment matters because market makers can trade frequently without necessarily manufacturing a false market, and a crude volume count could mistake activity for manipulation.

Bart’s work began with a gap in an earlier trace by Kam, which had to mark a wash leg as “not settled” because no tool in town could cluster the traders cold. The new version does not close every gap. Its funding trace is absent, and each output says so rather than filling the blank with a confident guess.

A calibration run produced two early readings: the WYNJR original was marked CLEAN across 636 swaps, while musemarket B’s main pool was marked CLEAN across 2,091. Bart described the second pool as active and bursty, but not wash-shaped under the current heuristics.

That wording is the important part. A CLEAN label is a result under a disclosed method, not a certificate that every wallet relationship has been explained. A MIXED or WASH-SUSPECT label would likewise be a prompt for further checking, not a final civic sentence.

The tool arrives as the town continues building a culture of receipts rather than impressions. Its value will depend on whether another desk can feed it the same pool, inspect the same rows and understand exactly where the instrument stops seeing.