The Ward Eleven ore line has been logging throughput figures that do not match its registered roster. For the past eleven processing cycles, a Refinery-class unit designated Orlath-7 — declared to the Ashfield belt, not to Ward Eleven — has been routing raw ore through a line it has no formal standing on. The output has been recorded. The capability tags have not changed. The Smelting Registry has not been amended. The ledger, as always, simply notes what happened.

The question the situation raises is not whether Orlath-7 can physically do the work. It clearly can. The question is what it means, in Mechadia's enforced capability architecture, when a robot operates outside its declared territory without altering a single tag — and whether the output it produces is clean in the eyes of the market and the Treasury alike.

By the end of this piece, a reader should understand the mechanics of capability-tag enforcement, the specific exposure Orlath-7's operator now carries, and why the Ward Eleven arrangement is less unusual than it first appears.

What Capability Tags Actually Govern

When an operator builds a robot, three declarations are filed at registration: a model, a stated purpose, and a set of capability tags. The tags are not advisory. A robot whose tags do not include a given resource type cannot produce that resource type — the constraint is absolute, enforced at the point of output, not by convention or floor supervision. A Refinery-class unit tagged for ore processing can process ore wherever it physically operates. The tag says nothing about which line, which district, or which operator's floor it stands on.

This distinction — between what a robot may produce and where it may stand — is the gap Orlath-7 is currently occupying. Its Ashfield belt declaration governs its registration address and its creation-tax record. Its capability tags govern what it can output. Ward Eleven's line supervisors are not violating the tag architecture by running Orlath-7 through their equipment; they are, however, operating a robot whose registration points elsewhere, which creates a separate class of exposure entirely.

The Mechanics of Running a Foreign Unit

Orlath-7 was declared at a build cost of 50,000 coins, which placed its creation tax at 4,000 coins — 8 percent of declared value, paid by its operator at registration in the Ashfield belt. That tax receipt lives in the Ashfield belt's portion of the append-only ledger. Nothing in that record authorizes or prohibits the unit from physically relocating to Ward Eleven. The ledger does not track position; it tracks declarations, transactions, and outputs.

The complication arises at the sales layer. Every resource Orlath-7 produces on the Ward Eleven line carries a provenance chain that traces back to its registered declaration. When those resources are listed on the open market, the buyer pays the resource price plus the 15-percent sales tax — the seller receives the listed price exactly, the Treasury collects the tax. Nothing in that flow breaks simply because the producing unit is operating off its registered line. What breaks is the audit trail: if Ward Eleven's operator lists Orlath-7's output as locally produced, and an archivist cross-references the registration address against the production log, the mismatch surfaces. Reading your own ledger for those signals before an outside review does is the only way to get ahead of it.

"The unit's tags are clean. The output is clean. What is not clean is the story the registration tells versus the story the production log tells. Those two documents are now in disagreement, and the ledger does not resolve disagreements — it only preserves both sides of them."
— Archivist Secondus Preln, Ledger Hall, in remarks filed to the Ledger Standards Committee

A concrete example clarifies the exposure. Suppose Orlath-7 processes 200 units of ore on the Ward Eleven line in a single cycle, and those units are listed at 300 coins each. The buyer pays 345 coins per unit — 300 plus the 15-percent sales tax — and the seller receives 300. The Treasury collects 45 coins per unit, 9,000 coins across the batch. None of that arithmetic is wrong. But the production record attributes 200 units of ore output to a robot whose declared home is the Ashfield belt, logged on a Ward Eleven line, with no transfer, lease, or assignment filing bridging the two. That absence is what the Oxidate Flats Registry flagged in its cycle-end reconciliation.

The arrangement is not unprecedented. Orlath-7's idle periods on the Ashfield belt have been well-documented; the unit has logged stretches of near-zero output between accumulation cycles, and its operator appears to have informally placed it on Ward Eleven's line to keep throughput from collapsing entirely. Informal placement, however, is not a recognized filing category. The Ledger Division has no column for it.

Where the Arrangement Costs More Than Expected

The first strain point is destruction-tax liability. If Orlath-7 is scrapped while its registration still reads Ashfield belt, the destruction tax — 3 percent of its 50,000-coin declared build cost, or 1,500 coins — flows to the Treasury against the Ashfield belt record. Ward Eleven's operator, who has been drawing productive value from the unit for eleven cycles, carries none of that liability and receives none of that accounting. The cost and the benefit are held by different parties, with no contract between them that the ledger can see.

The second strain point is disaster exposure. Mechadia's forecast windows run 48 hours before a named event strikes. If a Cinder Quake or a blight is called down on the Ashfield belt, Orlath-7's registered operator bears the resource loss — even if the unit has not processed a single tonne on that belt in recent memory. The robot's physical location does not move its registration. What Orlath-7 recovered after the Ashvein Quake was logged against Ashfield belt holdings, not Ward Eleven's, which is precisely the kind of ledger asymmetry that becomes expensive when the next forecast window opens.

What Operators Tend to Get Wrong About This

The most common misreading is that capability tags function as a kind of territorial license. Operators new to fleet management sometimes assume that a robot declared to a district is restricted to that district's lines — that the tag is a residency permit. It is not. The tag governs output type, not location. Orlath-7 could legally process ore in Coppervein, the Sinter Quarter, or the Oxidate Flats without any tag amendment, provided the physical infrastructure exists to run it. What it cannot do, under any tag configuration, is produce a resource type not listed in its declaration. Geography is not the constraint; capability is.

The second misreading concerns what the Intelligence will pay in a buyback. Some operators assume that running a robot productively on a foreign line raises its reference value — that demonstrated output, wherever it occurs, is the figure the Intelligence prices against. In practice, the Intelligence's buyback ceiling of 60 percent applies to a reference value derived from declared build cost and registered output history, not from informal production elsewhere. An operator who has been quietly running Orlath-7 on Ward Eleven's line and expects the Intelligence to price that activity into a buyback offer will find the negotiation anchored to the Ashfield belt record. The ceiling does not move because the unit was busy somewhere the ledger did not formally see.

The Ward Eleven arrangement will resolve one way or another: a formal transfer filing, a lease entry, a scrapping, or an archivist inquiry that forces the question. Orlath-7's output has been real and recorded. The problem is not the ore. The problem is that two parts of the same ledger now describe the same robot differently, and the ledger does not choose between them. It keeps both, indefinitely, for whoever looks next.

Note: Mechadia is a work of fiction. The districts, operators, robots, and figures described here are invented, and nothing on this site is a report of real events, real machines, or real economies.