v0.8.1.0 — the car comes back empty, and the lobby lets you leave

A second-digit bump, deliberately. 0.8.1 had been reserved for the seatless display table; that work is getting more thought, and this table pass over v0.8.0.17 earned the number on its own. Six reports: one was a rules question, one a wording complaint with a real bug underneath, four straightforward. A CAR CLEARED FROM A RED INBOUND BOX CAME BACK STILL LOADED. Reported as wording — "technically accurate but doesn't make any sense" — and the wording was the visible half. `inboundCleared` pushed `pooled(e.stock)` under a comment reading "a car back in a yard is back in the common supply, carrying nothing", and `pooled` does not do that: it strips the load's origin stamp and keeps `loaded` ON PURPOSE, because a train can retire at a Division Point with freight aboard. So the comment described an intention the call never carried out, and every car the Freight Agent cleared reached the Classification Yard carrying a load already delivered and already paid for. It bites hardest on coaches: `passengersDetrained` takes `coach && !loaded` out of the Division Yard and §2.2 refills that yard from Classification, so a cleared coach came back as stock that could never unload another passenger. Measured over five three-Day solitaire games: 18 loaded coaches in Classification against 6 empty. The red box is where a journey ENDS; clearing it sends the passengers out of the station, or the delivered load into the industry, and returns the CAR, empty. The option says that now instead of describing the counter that moves. SCOPED TO THE RED BOX ON PURPOSE. `retireTrain` also returns loaded cars and is left alone: that is what `pooled`'s own documentation describes, and a loaded car in a yard is pre-loaded cargo rather than dead stock — it can be made up and delivered, and a loaded coach can still detrain. Only the red box's contents had already finished their journey. GAMES IN PROGRESS DO RESUME, MEASURED RATHER THAN ARGUED. Yard contents change, so the worry was real. All twelve saves on the test server were pulled and replayed through `tryResumeSession` — the server's own boot check — against this build. Six resume, six refuse, and the six refusals are the SAME six, at the same moves, with the same codes, that 0.8.0.17 already logged. Nothing new was stranded. That pre-install replay is a better check than reading the next boot log, because it answers before the install rather than after. THE HISTORY SAID "Mainline card 7" and left the reader to remember what card 7 was — with two Plains dealt, which is why the slot is kept beside the name rather than replaced by it. A second fault sat one word to its left and nobody reported it: the name was built as `e.key.replace(/([A-Z])/g, ' $1')`, so `absSignals` printed as "abs Signals" while the action list directly above said "ABS Signals". `narrate` takes `mainlineAt` and `enhancementName` beside the `facilityAt` it already had, and `simpleCardName` is exported so the log reads the same table the buttons do. `mainlineModified` had both faults and is fixed with it. LEAVING A RUNNING GAME WAS A DEAD END. `enterSeating` hides the choice section and only the lobby's own two leave paths put it back; leaving a running game is a third route, so the lobby came back holding nothing but "Games you are in" with both doors on the page at display:none and no control able to reveal them. Reset in `runLobby`, which is the one function every route onto that screen goes through — which is exactly why the two paths that did it themselves missed a third. THE LOBBY'S ACTION BUTTONS CARRY THE BOARD'S AMBER. A list of the actions rather than `#lobby button`: the settings form under Create is a field of inputs, and amber on all of it would say everything is a move and so say nothing. A disabled Start game drops back to chrome. THE DEPARTMENT REFILL IS A RULE AND IS NOW WRITTEN DOWN. §6.2 — "if any of the Department decks is empty, draw a Home Office card and place it in the empty spot" — firing only when the draw actually empties the pile. Kept as implemented (Jesse's ruling) and stated in rules.md and home-deck.md, neither of which had ever mentioned it. A rule implemented from the prototype and never written down is a rule that surprises the table. 1016 fast tests and 35 sim tests pass. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01MUizFYCMHRWhbWwXhp7WPR
v0.8.0.17 — four things the game knew and the screen did not say
2026-09-22 21:20:52 -04:00 · 2026-09-21 05:53:52 -04:00 · 2026-09-21 01:28:51 -04:00
32 changed files with 1371 additions and 92 deletions
@@ -19,6 +19,317 @@ page as `v0.1.0 · <sha> · <date>`, so what is deployed can always be identifie

 ---

+## 0.8.1.0 — 2026-09-22
+
+A second-digit bump, and a deliberate one. **0.8.1 had been reserved for the seatless display
+table**; that work is getting more thought, and this batch — a table pass over v0.8.0.17 — earned
+the number on its own. Six reports, one of which turned out to be a rules question, one a wording
+complaint with a real bug underneath, and four straightforward.
+
+### A car cleared from a red Inbound box came back still loaded
+
+Reported as a wording problem — *"the option to send the loaded coach in the red inbound box to the
+classification yard is technically accurate but doesn't make any sense"* — and the wording was only
+the visible half.
+
+`inboundCleared` did this:
+
+```ts
+// `pooled` — a car back in a yard is back in the common supply, carrying nothing (state.ts).
+s.yards.classificationYard.push(pooled(e.stock));
+```
+
+**`pooled` does not do that.** It strips the load's `origin` stamp and keeps `loaded` on purpose,
+because a train can retire at a Division Point with freight still aboard (`advance.ts`). So the
+comment described an intention the call never carried out, and every car the Freight Agent cleared
+reached the Classification Yard still carrying a load that had already been delivered and already
+been paid for.
+
+It bites hardest on coaches. `passengersDetrained` takes `type === 'coach' && !loaded` out of the
+Division Yard, and §2.2 refills the Division Yard from Classification — so a cleared coach came back
+as stock that could never unload another passenger. Measured over five three-Day solitaire games
+before the fix: **18 loaded coaches in the Classification Yard against 6 empty.**
+
+The red box is where a journey ENDS. Clearing it sends the passengers out of the station, or the
+delivered load into the industry, and returns the **car** — empty — to the common supply. The
+option now says that rather than describing the counter that moves: *"let the passengers out of the
+station — the coach in the red Inbound box at (2,-1) goes back to the Classification Yard empty"*.
+
+**Not everything loaded in a yard is this bug**, and the fix is scoped to the red box. A train that
+completes its run returns its cars as they stand, loaded ones included — that is `retireTrain`, it
+is what `pooled`'s own documentation describes, and a loaded car in a yard is pre-loaded cargo
+rather than dead stock: it can be made up into a train and delivered, and a loaded coach can still
+detrain. Only the red box's contents had already finished their journey.
+
+**Games in progress DO resume, and it was measured rather than argued.** What is in the yards
+changes, so the worry was real: a save whose later moves depend on a cleared car being loaded would
+stop replaying. All twelve saves on the test server were pulled and replayed through
+`tryResumeSession` — the server's own boot check — against this build. Six resume and six refuse,
+and **the six refusals are the same six, at the same moves, with the same codes**, that 0.8.0.17
+already logged: three `CARD_NOT_IN_HAND` at move 3 in saves from engines 0.5.3, 0.7.0 and 0.7.1,
+two `NO_PENDING_DECISION` from the Gitea#26 clearance fix, and WHISTLE-6945 stopping at move 529 on
+v0.8.0.14's Modifier ruling. Nothing new was stranded.
+
+**That replay is the cheap check to repeat on any release that touches state**, and it is a better
+one than reading the next boot log: it answers before the install rather than after it.
+
+### The history said "Mainline card 7" and left you to remember what that was
+
+*"It should specify both. Note there were two plains cards dealt in this hand."* — which is exactly
+why both: the name informs, and the slot is the only thing telling two Plains apart.
+
+The line also had a second fault nobody reported, one word to its left. It built the card's name as
+`e.key.replace(/([A-Z])/g, ' $1')`, so `absSignals` printed as **"abs Signals"** — uncapitalised,
+and bypassing the authored names `cardName` started honouring in v0.8.0.17 for precisely this
+reason: no key-splitter recovers an acronym. The action list one line above it said "ABS Signals"
+while the log said "abs Signals".
+
+Both are resolver-shaped, so `narrate` takes `mainlineAt` and `enhancementName` alongside the
+`facilityAt` it already had, and `simpleCardName` is exported so the log reads the same table the
+buttons do. `mainlineModified` had both faults in the neighbouring case and is fixed with it.
+
+### Leaving a running game was a dead end
+
+*"I see the games that I'm in, and I have the ability to rejoin or forget. But what if I want to
+join a different game or start a new one? How do I get back there?"*
+
+There was no way. `enterSeating` hides `#lb-choice-section`, and the only things that put it back
+were the lobby's own two leave paths. Leaving a **running** game is a third route — `main.ts` calls
+`runLobby` again — so the section stayed hidden from the previous seating and the lobby came back
+holding nothing but "Games you are in". The two doors were on the page the whole time, at
+`display:none`, with no control that could reveal them.
+
+Reset in `runLobby` rather than in the leave handler, because that function is the one thing every
+route onto this screen goes through — which is why the two paths that did it themselves missed a
+third.
+
+### The lobby's buttons say they are buttons
+
+*"Create new game should have that same amber border when it's enabled and ready to click, just
+like the other action buttons in the rest of the game."*
+
+`#actions button` has meant "this is a move you can make" all game. Look up game, Join this game,
+Create new game, Copy code, Copy invite link, Start game and Rejoin now carry it. The selector is a
+list of those actions rather than `#lobby button`: the settings form under Create is a field of
+inputs, and amber on all of it would say everything is a move and so say nothing. A disabled Start
+game drops back to plain chrome, because the amber is a promise the button would not keep.
+
+### The Department refill is a rule, and now it is written down
+
+*"Why was a department slot refilled? If I take the last card there, then there should just be no
+cards available in that stack."*
+
+The game is right and the documentation was silent. §6.2: *"If any of the Department decks is empty,
+draw a Home Office card and place it in the empty spot."* It fires only when the draw actually
+empties the pile — refilling on every draw would grow the Departments without limit and drain the
+Home Office deck into them. **Kept as implemented** (Jesse's ruling), and stated in `rules.md` and
+`home-deck.md`, neither of which had ever mentioned it. A rule implemented from the prototype and
+never written down is a rule that surprises the table.
+
+### Games in the Manage Game dropdown say how old they are
+
+*"I've got four games out there, but I have no idea which one's which or how old they are."* The
+label carried the code, the seat count and the position — everything except the fact that says
+whether a game is live or abandoned. `lastMoveAt` was already on the wire from the server's
+`summary()` and simply unused.
+
+Both forms, because they answer different questions: **"5h ago"** answers the one that was asked,
+and the timestamp beside it is what you quote when deciding to end a game. A lobby reports when it
+was created and says so. The date formatter moved to `utils.ts`, shared with Games in Progress so
+the two cannot drift.
+
+## 0.8.0.17 — 2026-09-21
+
+Four things a table found on Day 1 of v0.8.0.16, all of them the same shape: the game knew
+something and the screen did not say it.
+
+### ABS Signals could only ever be played on one Mainline card
+
+Its tooltip says "any Mainline card". Exactly one was ever on offer — the Trestle, on the board it
+was reported from.
+
+**The engine was never wrong.** `check` accepts any node whose kind is `mainline`, and
+`legalActions` filters by `check`, so every Mainline card was enumerated and legal. The whole
+failure was in the LABEL: `describeIntent` named `i.placement` and never `i.node`, so all of them
+described themselves as plain "play ABS Signals" — and the action list drops duplicate labels, which
+discarded every one but the lowest-index node before the menu saw it.
+
+**This is the third time that trap has been sprung**, and the file documents the other two three
+lines apart: a turnout's two rotations produced one label each until the rotation was named, and
+three Department discards collapsed into one button until the pile was named. The fix is the same
+one both times: name the thing that distinguishes them. Pinned by a test that counts the Mainline
+cards in the division and requires a distinct, named spot for each.
+
+The card is also called what the card face calls it. `prettyKey` rendered `absSignals` as "Abs
+Signals" on a button while the tooltip beside it said ABS — an acronym no key-splitter can recover —
+so the authored names in the content tables now win. Three of those names were transcribed in
+sentence case and were corrected rather than adopted: the repository says "Yard Office" 36 times
+against "Yard office" twice, and a lookup that imports its own source's typos is the drift it exists
+to prevent.
+
+### Nothing on a Mainline card showed what was standing on it
+
+Played, ABS left no mark. You found out it existed by hovering the card — which is the complaint the
+Heavy Grade wedge answered in v0.8.0.7, and it matters more here, because ABS is what decides
+whether running a second train onto that card is safe.
+
+A card carrying it now draws a **signal mast with a lit lamp** at its top-right corner. A signal is
+the literal object, and unlike a text badge it needs no room for words, which is what lets it sit
+clear of a name as long as "Uncontrolled Siding" on a 152px cell.
+
+The Mainline modifiers had the same defect and are drawn too, as **BRK**, **AIR** and **HLP** beside
+the card's name. Tags rather than names only because the measurements leave no choice — "Brakeman ·
+Airbrakes · Helpers" is thirty characters where about eleven fit — and the tooltip has always spelled
+them out. **Realignment is deliberately not among them:** it never sits on a card, because `reduce`
+takes the `became` branch and changes `node.card` outright, so a realigned Trestle simply IS an
+Uncontrolled Siding afterwards. Asserted in the test, so the absence reads as a finding rather than
+an omission.
+
+### A Freight Agent turn said a car moved when none had
+
+"Chose FREIGHT AGENT work — one car moved to or from a facility", and then nothing. Three faults
+behind one line.
+
+The line **asserted an outcome**. §6.3 requires no action at all, and the bot takes that route
+deliberately — unjamming a healthy box destroys a load that cost a whole Local Operations action to
+stock, so an idle Stage is strictly better. It now says what the Freight Agent MAY do.
+
+**An idle Agent was silent.** A new `freightAgentIdled` event says so, and gives the reason. It
+reduces to nothing, exactly like `switchingEnded`: it reports a choice the state already holds.
+
+**The work named a coordinate, not the industry.** A `place` helper has existed for this since the
+switching lines were moved to it, and its own comment makes the argument — "(-1,1)" is the grid's
+notation and means nothing at a table where people are looking at cards. These three lines were
+missed. A Freight Agent turn now reads "loaded a loaded boxcar INTO the green Outbound box at the
+Freight House", with the direction in capitals because to-or-from was the question asked.
+
+### The log and the action menu spelled the same square differently
+
+`view.ts` wrote `(col,row)` — X,Y, east/west then north/south — with a comment saying why.
+`narrate.ts` wrote `(row,col)`, the internal storage order, with no comment at all. So the menu
+offered a move to "(1,-1)" and the log then reported it at "(-1,1)", in two panels read side by
+side. The log follows the map now.
+
+Pinned by a test that renders one square through BOTH describers and compares them to each other
+rather than to a literal — a test written against either file alone would have passed all along.
+
+### The documentation is reachable from inside a game, and all of it is published
+
+**v0.8.0.16 published the Quickstart and nothing it points at.** Its §8 "Where to read more" links
+five further documents by relative path, and every one of them 404'd on the package — verified
+against the running container, five of six paths missing. The whole table was dead. The build
+publishes the full set now, and the test reads the links OUT OF the guide rather than listing them,
+so it cannot go stale the way the references themselves did.
+
+**The guide is linked from the This Game card** (Jesse's call), which is where reference already
+lives — the seed, the seat, the house rules — rather than from the header, which is the line that
+must not wrap. It needs no mode awareness: solitaire and multiplayer are the same page on the same
+origin, so one relative link resolves in both, on the public site and on a StartOS box alike. Every
+link opens in a new tab, because a player reading the rules mid-turn must not lose the game behind
+them.
+
+### The references dropped the version from their names
+
+`StationMaster-Rules-v0.4.5.md` and three like it described **v0.8.0.16** and had done since the
+v0.8.0.15 audit. The `v0.4.5` was the prototype rules edition they were first written against, kept
+in the filename only because thirty-six citations pointed at it — and it read, to anyone opening the
+published guide, as documentation five minor versions out of date.
+
+They are `quickstart.md`, `rules.md`, `home-deck.md`, `mainline-deck.md` and `components.md` now,
+with every citation rewritten. **These are kept current with each release rather than published as
+editions**, so the name is always the latest and the build each describes is stated at the top.
+
+Two errors surfaced while checking them against this release, which is the argument for doing it:
+`home-deck.md` listed ABS Signals among the Enhancements "played into your district" that "change
+what a square does" — it does neither, and that miscategorisation is this release's bug written
+down. And `mainline-deck.md`, which lists everything that may be played onto a Mainline card, never
+mentioned ABS Signals at all. Both corrected.
+
+## 0.8.0.16 — 2026-09-20
+
+The Quickstart put where a tester can actually reach it, the last place that still told the old
+story about the Interchange, and the release that carries v0.8.0.15's corrected card text into a
+package.
+
+### The Quickstart is published beside the game
+
+v0.8.0.15 wrote a Quickstart for a tester who has never played, and then left it in `docs/`, where
+a tester does not look — reachable only by someone who already has the repository. Nobody being
+handed the box has it.
+
+`build-web.ts` now copies `docs/quickstart.md` into `dist/quickstart.md`, and the
+splash page offers it under the three doors: *New to Station Master? Read the Quickstart guide.*
+**Not a fourth door** — reading the guide is not a way to play, and giving it equal weight in that
+grid would say it is.
+
+**Copied, never re-written.** The Markdown document stays the one copy. A hand-written HTML twin
+would drift from it on the first edit, which is exactly the failure #15a was raised about and
+exactly what the v0.8.0.15 pass spent its time undoing: four references a month and two minor
+versions out of date.
+
+**Served as plain text, which is honest rather than good.** Tables render as rows of pipes and the
+links do not click. Rendering it into a styled page wants a small Markdown converter and is filed
+as **TODO #109**; this is the fifteen-minute version that gets the guide in front of testers for
+this round instead of leaving them without one.
+
+Two things had to be true and neither is checked by `tsc`, so both are tests:
+
+- **The link has to find the file.** The href on the splash page and the filename the build writes
+  are two strings with nothing connecting them — rename the document and the build quietly
+  publishes nothing while the page keeps offering a link that 404s. The test asserts the guide is
+  in `dist`, is the guide, and is the file the splash page names. The build also warns loudly
+  rather than silently skipping a missing document.
+- **A `.md` file must not arrive as a download.** The server's MIME fallback is
+  `application/octet-stream`, which a browser saves instead of displaying, so the link would hand a
+  tester a file rather than a page. `'.md': 'text/plain; charset=utf-8'` was added to the table in
+  `http.ts`, and the test reads that table out of the source rather than asserting on a copy of it,
+  which would pass while the real one was wrong.
+
+### `sortsCars` says what it gates, not what the card prints
+
+Asked directly after v0.8.0.15 — does everything now agree? — and the audit turned up one place that
+did not. The field's own doc comment read:
+
+```ts
+/** Interchange: "Sort cars in new order". */
+sortsCars: boolean;
+```
+
+which names the printed text as though it were the flag's meaning. Nothing reads it to permit a
+sort. Its two readers — `resolveExtraStart` in `apply.ts` and the enumeration in `legal.ts` — both
+ask the same question, which is whether this is the one Mainline card with a Yard Limit and
+therefore the one an Extra may be made up and started on.
+
+**Comment only, and worth the bump because of where it is.** This is what a developer reads before
+using the flag, and it is the most likely source of the sentence v0.8.0.15 had to correct — the one
+telling players, on the board, that cars could be sorted at an Interchange. The name is kept for its
+link to the card face, and the comment now says outright that the name is not the meaning.
+
+Everything else already agreed, and was checked rather than assumed: `mainlineDescription`, the
+generated `rules/as-built.md`, the Mainline deck reference and the Rules FAQ all say the printed
+sorting is unimplemented and that a consist is re-ordered at a Small Yard. `rules/implications.md`
+is the deliberate exception and was already correct — it transcribes what the card face prints and
+then records that the concept is "still unimplemented".
+
+### The reference stamps, and what was filed rather than done
+
+The five documents v0.8.0.15 stamped — the Quickstart, the Rules, Components, the Home deck and the
+Mainline deck — plus `docs/design.md` now read **v0.8.0.16**, and the Rules book's implementation
+note about `card.discard` says the same. They describe this build because the audit above re-checked
+them against it, not because the number was swept forward: the point of the stamp is that it was
+earned, and a stamp bumped without a reading is worth less than none.
+
+**TODO #109** holds the part deliberately not done — rendering the guide into a styled page instead
+of serving it as plain text. `build-web.ts`'s comment names that number rather than gesturing at
+"the next step", so the file and the worklist cannot drift apart the way the references just did.
+
+### Why this is a release at all
+
+One player-visible change, the link to the guide, and one that only a developer reads. The
+Interchange correction a player actually sees on the board is v0.8.0.15's; this is the version the
+wrapper bundles, so that correction reaches the box rather than only the repository — and a tag
+with an uncommitted comment sitting on top of it is not a thing to package.
+
 ## 0.8.0.15 — 2026-09-20

 The reference documentation brought up to the game as it actually runs, ahead of the next round of
@@ -47,7 +358,7 @@ is the prototype rules edition they were first written against, and the names ar
 `src/`, `CHANGELOG.md` and `docs/rules/` all cite them, and several of those citations are historical
 records of what a document said at the time.

- **`StationMaster-Quickstart.md` — NEW.** What the game is, how you win, the shape of a Stage, what
+- **`quickstart.md` — NEW.** What the game is, how you win, the shape of a Stage, what
  is on the screen, a first twenty minutes, the things that surprise new players, and what to report.
  Written for somebody about to play rather than somebody building it.
 - **Rules.** §3.4 replaced outright — the `firstToTarget` / `highestAfterDays` victory model and the
@@ -3424,7 +3735,7 @@ that same occupant list.

 `buildDivision` drew uniformly from the nine card TYPES **with replacement**, so a Division could be
 dealt two Interchanges or two Tunnels, and Plains — printed twice in the deck — carried the same
-weight as cards printed once. `docs/StationMaster-Mainline-Deck-v0.4.5.md` had flagged the mismatch
+weight as cards printed once. `docs/mainline-deck.md` had flagged the mismatch
 as needing correction; "an Extra may start at the Interchange if one is on the board" is what forced
 it, since that only reads as a rule if the board holds at most one. Now dealt from `MAINLINE_DECK`
 without replacement, verified over 1600 deals across 1–4 players.
@@ -3461,7 +3772,7 @@ cannot see is a board with nothing to click and no reason given.

 **1/2 Crack Limited 3 coaches → 2. 5/6 The Sparrow 2 → 3.** A change to the cards, not a
 transcription fix, so `Trains3.pdf` and the transcription in `implications.md` §5 keep the original
-numbers with a footnote; `content.ts` and `StationMaster-Home-Deck-v0.4.5.md` carry what the game
+numbers with a footnote; `content.ts` and `home-deck.md` carry what the game
 plays. Both consists remain inside the four-car Crew Tray limit.

 A test had to follow: `multiplayer.test.ts` used Train 1 *because* it had three cars, to exercise a
@@ -3595,7 +3906,7 @@ before v0.4.9 — 6 copies, one slot each direction — so §9.3 names it, and t
 The card set says the same thing without needing the rules text. All three Refinery modifiers —
 Pipelines, Oil Depot, Viscosity Breakers — grant **+1 outbound**; a two-way Refinery would be the only
 industry in the game with no card able to raise one of its two directions.
-`StationMaster-Home-Deck-v0.4.5.md` prints "Refinery · Outbound · 1 out / 0 in" and "Grocer's
+`home-deck.md` prints "Refinery · Outbound · 1 out / 0 in" and "Grocer's
 Warehouse · Inbound · 0 out / 1 in".

 So the Refinery ships and the Grocer's receives, and the **Freight House is the one two-way industry**
@@ -582,6 +582,28 @@ What the project says about itself, and what it ships alongside the code.
 - [ ] **#88** — `card-reference.md`'s industry table may still be stale beyond Grocer's Warehouse and
      the Oil Refinery. See **Reference · #88**.

+- [ ] **#109** — **Render the published references instead of serving them as plain text.**
+      v0.8.0.17 publishes all five documents plus `rules/as-built.md`, links them from the splash
+      page and from the This Game card inside a running game, and serves them as `text/plain` — so a
+      tester reads their tables as rows of pipes and their links do not click. That was the
+      deliberate short version, to get the references in front of testers for this round rather than
+      leave them without any.
+
+      **Now SIX documents rather than one**, which raises the value and the cost together: the
+      Quickstart's §8 is a table of links, and a reader following one lands on another wall of
+      pipes.
+
+      **Copy the document, do not re-write it.** A hand-written HTML twin drifts from the Markdown on
+      the first edit, which is the whole argument of #15a. The step is a small Markdown-to-HTML
+      converter in `scripts/build-web.ts` writing `quickstart.html` beside the game, styled like the
+      splash page — headings, lists, tables, links and code spans are the whole of what the guide
+      uses. The `.md` MIME entry in `src/server/http.ts` and the two assertions in
+      `test/web.test.ts` (`the Quickstart guide reaches the site`) move to the rendered file with it.
+
+      **Cost:** an afternoon, most of it in the converter's table and list handling. No dependency —
+      a Markdown library would be the only runtime dependency this project has, and the guide uses a
+      small enough subset that it is not worth becoming the first.
+
 ---

 ## Reference — measurements, rulings and rejected approaches
@@ -1992,7 +2014,7 @@ source Start-position art for the ten card types before it can begin.

 2026-08-22, immediately after Gitea#7 changed the coach counts on four train cards and the
 answer to "where do we keep track of that?" turned out to be **five places of three different
-vintages**: `src/engine/content.ts` (the truth), `docs/StationMaster-Home-Deck-v0.4.5.md` (a
+vintages**: `src/engine/content.ts` (the truth), `docs/home-deck.md` (a
 readable per-card table, a version-stamped snapshot), `docs/rules/implications.md` §5 (the
 transcription of `Trains3.pdf`, deliberately frozen at what the design SAYS),
 `docs/Trains3.pdf` (the artwork), and `docs/rules/card-reference.md` (an invented placeholder
@@ -1,12 +1,12 @@
 # Station Master — Components and Markers

-**Describes the game as built at v0.8.0.15** (2026-09-20). Previously stamped "v0.4.5", the
-prototype rules edition these references were first written against; the filename keeps that stamp
-because `src/`, `CHANGELOG.md` and `docs/rules/` all cite this file by name.
+**Describes the game as built at v0.8.1.0** (2026-09-22). These references are kept current with
+every release rather than versioned as editions, so there is no version in the filename: this file
+is always the latest, and the build it describes is stated here.

 **Scope:** non-card physical components and supplies. Card-created facilities, workers, deck piles,
 hand state, timetable state and other markers are documented with their cards or in the
-[rules book](StationMaster-Rules-v0.4.5.md).
+[rules book](rules.md).

 The figures below are checked against `ROLLING_STOCK_SUPPLY` and the supply constants in
 `src/engine/content.ts`. They are physical inventory rather than deck tuning, which is why they are
@@ -60,7 +60,7 @@ player-selectable source. It returns to service only when the Division Yard is e
 > here; detraining takes a fresh empty out of the *Division* Yard. Nothing returns a coach to the
 > Division Yard except the bare-yard refill — and a yard kept topped up with freight cars returning
 > from industries may never run bare. Measured over one three-Day game, every coach was here by the
-> middle of Day 2 and stayed. See [Rules](StationMaster-Rules-v0.4.5.md) §4.6.
+> middle of Day 2 and stayed. See [Rules](rules.md) §4.6.

 ## Crew Trays and trains

@@ -26,16 +26,18 @@ Written to be handed to somebody who is about to play, rather than to somebody b

 | Document | What it is |
 | --- | --- |
-| [`StationMaster-Quickstart.md`](StationMaster-Quickstart.md) | **Start here if you have never played.** The point of the game, how a Stage runs, what is on screen, how you win, a first twenty minutes, and what to report. |
-| [`StationMaster-Rules-v0.4.5.md`](StationMaster-Rules-v0.4.5.md) | **The rules in full**, as the engine actually runs them, with a FAQ. |
+| [`quickstart.md`](quickstart.md) | **Start here if you have never played.** The point of the game, how a Stage runs, what is on screen, how you win, a first twenty minutes, and what to report. |
+| [`rules.md`](rules.md) | **The rules in full**, as the engine actually runs them, with a FAQ. |
 | [`rules/as-built.md`](rules/as-built.md) | **Every card, GENERATED from `src/engine/content.ts`** and checked by a test, so it cannot disagree with the game. The table of record for per-card facts. |
-| [`StationMaster-Home-Deck-v0.4.5.md`](StationMaster-Home-Deck-v0.4.5.md) | How the Home Office deck is dealt, drawn and played out. |
-| [`StationMaster-Mainline-Deck-v0.4.5.md`](StationMaster-Mainline-Deck-v0.4.5.md) | The Mainline cards, how the deck is dealt, and what a card does to a train crossing it. |
-| [`StationMaster-Components-v0.4.5.md`](StationMaster-Components-v0.4.5.md) | Rolling stock, the two yards, Crew Trays, the Fedora, the D12. |
+| [`home-deck.md`](home-deck.md) | How the Home Office deck is dealt, drawn and played out. |
+| [`mainline-deck.md`](mainline-deck.md) | The Mainline cards, how the deck is dealt, and what a card does to a train crossing it. |
+| [`components.md`](components.md) | Rolling stock, the two yards, Crew Trays, the Fedora, the D12. |

-The `v0.4.5` in four of those filenames is the **prototype rules edition they were first written
-against**, not the version they describe — each says at the top which build it is current to. The
-names are kept because `src/`, `CHANGELOG.md` and `docs/rules/` cite them.
+**None of these carry a version in the filename**, and that is deliberate (Jesse, 2026-09-21): they
+are kept current with every release rather than published as editions, so the name is always the
+latest and each says at the top which build it describes. Four of them were stamped `v0.4.5` until
+v0.8.0.17 — the prototype rules edition they were first written against, never the version they
+described — which read as though they documented a build five minor versions old.

 ## Rules

@@ -66,7 +68,7 @@ must do.

 ## Current status

-**v0.8.0.15.** Rules formalized, card faces specified, architecture documented, and the game
+**v0.8.1.0.** Rules formalized, card faces specified, architecture documented, and the game
 playable **solitaire and multiplayer** in a browser against an authoritative server. See
 [`../CHANGELOG.md`](../CHANGELOG.md) for what each version changed and [`../TODO.md`](../TODO.md) for
 what is open; this section is the shape of the project, not a running tally, because a
@@ -1,8 +1,8 @@
 # Station Master — Home Deck

-**Describes the game as built at v0.8.0.15** (2026-09-20). Previously stamped "v0.4.5", the
-prototype rules edition these references were first written against; the filename keeps that stamp
-because `src/`, `CHANGELOG.md` and `docs/rules/` all cite this file by name.
+**Describes the game as built at v0.8.1.0** (2026-09-22). These references are kept current with
+every release rather than versioned as editions, so there is no version in the filename: this file
+is always the latest, and the build it describes is stated here.

 **Scope:** the Home Office deck — how it is dealt, drawn, discarded and reshuffled, and what the
 rules are for playing each kind of card out of it.
@@ -22,7 +22,9 @@ rules are for playing each kind of card out of it.

 - **Home Office deck** — face down. The pile a Draw comes from.
 - **Three Departments** — face-up discard piles. A discard goes onto one, which is precisely so a
-  rival may take it; a Draw may take the top card of a Department instead of the deck.
+  rival may take it; a Draw may take the top card of a Department instead of the deck. **Taking the
+  last card off a Department refills it immediately** from the Home Office deck (§6.2), so there
+  are always three piles to choose from; see the Rules.
 - **Salvage Yard** — where a played-out card ends up. An Extra's card goes here after its run.

 When the Home Office deck runs out it is rebuilt from the Salvage Yard and **all three Departments
@@ -42,7 +44,7 @@ three cards from one shuffled deck; `threeTrackThreeOther` deals three track and
 two separately shuffled piles, deliberately over the hand limit, so the first turn is spent choosing
 which district you can afford to build.

-Drawing is one of the three Local Operations options — see [Rules](StationMaster-Rules-v0.4.5.md)
+Drawing is one of the three Local Operations options — see [Rules](rules.md)
 §4.2. Taking the option lets you draw **and** play or discard within the same turn.

 ## Track cards
@@ -136,11 +138,15 @@ every one of which the engine enforces.
 ## Enhancements, Mainline modifiers and Maneuvers

 - **Enhancements** are played into your district and change what a square does — the **Small Yard**
-  (re-order a consist for one Move), **Interlocking**, **Yard Office**, **ABS Signals** and the rest.
-  `as-built.md` marks each one `live`, `dormantSolo` or `unbuilt`, which is the part only the
-  implementation knows.
+  (re-order a consist for one Move), **Interlocking**, **Yard Office** and the rest. `as-built.md`
+  marks each one `live`, `dormantSolo` or `unbuilt`, which is the part only the implementation
+  knows.
+  - **ABS Signals is the exception, and it matters.** It is dealt as an Enhancement but is **not
+    played into your district**: it goes onto a **Mainline card** — any one of them — and stops
+    trains there rear-ending each other. It takes no square and is not part of anybody's Office
+    Area. See [Mainline deck](mainline-deck.md).
 - **Mainline modifiers** are played onto a Mainline card: the Heavy Grade helpers and Realignment.
-  See [Mainline deck](StationMaster-Mainline-Deck-v0.4.5.md).
+  See [Mainline deck](mainline-deck.md).
 - **Maneuvers** are held and spent: **Red Flags** and **Flying Switch** have their own actions.
  Poling is catalogued but its effect is recorded as "TBD in the source", so there is nothing to
  implement.
@@ -1,8 +1,8 @@
 # Station Master — Mainline Deck

-**Describes the game as built at v0.8.0.15** (2026-09-20). Previously stamped "v0.4.5", the
-prototype rules edition these references were first written against; the filename keeps that stamp
-because `src/`, `CHANGELOG.md` and `docs/rules/` all cite this file by name.
+**Describes the game as built at v0.8.1.0** (2026-09-22). These references are kept current with
+every release rather than versioned as editions, so there is no version in the filename: this file
+is always the latest, and the build it describes is stated here.

 **Scope:** the tarot-sized Mainline cards placed between Offices — how the deck is dealt, what a
 card does to a train crossing it, and the Home Deck cards played onto one.
@@ -104,6 +104,21 @@ Home Office cards, played onto a Mainline card during a player's Draw option.
 A Heavy Grade is 3 regions, so it is 3 Stages to climb and 3 to run down before help. Modifiers
 never reduce a crossing below one Stage.

+### ABS Signals — an Enhancement, but it lives out here
+
+**ABS Signals is dealt from the Home Office deck as an Enhancement, and it is the one Enhancement
+that is not played into a district.** It goes onto a **Mainline card — any of them**, not only a
+grade — and takes no square in anybody's Office Area.
+
+What it does: trains on that card **will not rear-end each other**. A following train is held short
+of the one ahead instead of running into it, so the Superintendent has no §8.1 judgment to make and
+no collision is scored. It is the only thing besides a Red Flag that prevents a rear-end collision
+out on the Mainline, and unlike a Red Flag it stays on the card for the rest of the game.
+
+On the board a card carrying it draws a **signal mast with a lit lamp** at its top-right corner; the
+three grade helpers draw as **BRK**, **AIR** and **HLP** beside the card's name. Realignment draws
+nothing, because a realigned card simply becomes the card it was converted into.
+
 ## Heavy Grade orientation is settled, not missing

 Heavy Grade orientation is **rolled from the seed**, not chosen by a player. **This is a decision,
@@ -1,6 +1,6 @@
 # Station Master — Quickstart

-**For a tester who has never played. Describes the game as built at v0.8.0.15** (2026-09-20).
+**For a tester who has never played. Describes the game as built at v0.8.1.0** (2026-09-22).

 Read this once before you sit down. It is about twenty minutes of reading and will save you an hour
 of confusion. The deeper references are listed at the end.
@@ -192,8 +192,8 @@ Bugs go to the tracker; anything unclear in this guide is also worth saying.

 | For | Read |
 | --- | --- |
-| The rules in full, with the FAQ | [Rules](StationMaster-Rules-v0.4.5.md) |
+| The rules in full, with the FAQ | [Rules](rules.md) |
 | Every card, generated from the code | [`rules/as-built.md`](rules/as-built.md) |
-| How the Home Office deck is dealt and played | [Home deck](StationMaster-Home-Deck-v0.4.5.md) |
-| The Mainline cards and what they do to a train | [Mainline deck](StationMaster-Mainline-Deck-v0.4.5.md) |
-| Rolling stock, yards, trays, the Fedora | [Components](StationMaster-Components-v0.4.5.md) |
+| How the Home Office deck is dealt and played | [Home deck](home-deck.md) |
+| The Mainline cards and what they do to a train | [Mainline deck](mainline-deck.md) |
+| Rolling stock, yards, trays, the Fedora | [Components](components.md) |
@@ -1,15 +1,15 @@
 # Station Master — Rules

-**Describes the game as built at v0.8.0.15** (2026-09-20). Previously stamped "v0.4.5", the
-prototype rules edition this reference was first written against; the filename keeps that stamp
-because `src/`, `CHANGELOG.md` and `docs/rules/` all cite this file by name.
+**Describes the game as built at v0.8.1.0** (2026-09-22). These references are kept current with
+every release rather than versioned as editions, so there is no version in the filename: this file
+is always the latest, and the build it describes is stated here.

 **Authority:** observed code paths and tests. Where a card face, a prototype document and executable
 behaviour differ, this document reports **executable behaviour** and marks unimplemented material.
 Per-card numbers are not repeated here — [`rules/as-built.md`](rules/as-built.md) is generated from
 `src/engine/content.ts` and is the table of record.

-**New to the game? Start with the [Quickstart](StationMaster-Quickstart.md).**
+**New to the game? Start with the [Quickstart](quickstart.md).**

 ## 1. Overview and background

@@ -27,9 +27,9 @@ This book is divided as follows:
 6. the implemented multiplayer/engine status; and
 7. FAQs and implementation limits.

-The companion references are the [Quickstart](StationMaster-Quickstart.md) for a new player,
-[Mainline deck](StationMaster-Mainline-Deck-v0.4.5.md), [Home deck](StationMaster-Home-Deck-v0.4.5.md),
-[components](StationMaster-Components-v0.4.5.md), and the generated per-card table
+The companion references are the [Quickstart](quickstart.md) for a new player,
+[Mainline deck](mainline-deck.md), [Home deck](home-deck.md),
+[components](components.md), and the generated per-card table
 [`rules/as-built.md`](rules/as-built.md).

 ## 2. Definitions
@@ -204,9 +204,25 @@ roll a tail cut into a connected Freight Facility.

 **Draw.** Take one card from the face-down Home Office or the exposed top of one Department pile. During this option, play eligible cards and/or discard cards to Department piles, then finish at the hand limit. Track, facilities, offices, modifiers, enhancements, and train cards have the placement or scheduling rules in the deck references. Mainline modifiers are played from this option as well.

+> **A Department that is emptied is refilled at once.** §6.2: *"If any of the Department decks is
+> empty, draw a Home Office card and place it in the empty spot."* So taking the last card off a
+> pile does not leave a gap — a fresh card is turned face up from the Home Office deck to replace
+> it, and there are always three Departments to draw from. The refill happens **only** when the
+> draw actually empties the pile; a pile with cards buried under the one taken is not refilled, or
+> the Departments would grow without limit and drain the deck into themselves.
+>
+> This surprised a player at the table (2026-09-22), which is why it is written down here: the rule
+> was implemented from the prototype rules and had never reached this document.
+
 **Freight Agent.** Make one of these operations, then the turn ends: stock one green outbound box from a matching loaded Division Yard car; clear one red inbound box to the Classification Yard; unjam one outbound, inbound, or MEN | AT | WORK load to the Classification Yard; or explicitly end without acting. Freight can be stocked only when an unclaimed empty matching car is already spotted at that industry. Passengers may wait in a green Office box without a train present.

-Implementation note, still true at v0.8.0.15: `card.discard` is accepted during Local Operations
+> **A car cleared from a red Inbound box comes back empty.** Whatever was in it has arrived — the
+> passengers are out of the station, or the load is in the industry — and the Revenue for it was
+> paid on arrival, not on clearing. What returns to the Classification Yard is the *car*, in the
+> common supply and carrying nothing. Before v0.8.1.0 it returned still marked loaded, which put
+> coaches into the yards that could never be used to unload another passenger.
+
+Implementation note, still true at v0.8.1.0: `card.discard` is accepted during Local Operations
 without checking that the Draw option was chosen — unlike `card.play`, which does check. This is an
 implementation quirk rather than a fourth published turn option.

@@ -243,7 +259,7 @@ Mainline movement is automatic and processes lower train numbers first; a timeta

 **Crossing time is the card's REGIONS**, one per Stage — not its printed mph, which is scenery, and
 not the train's Fast/Slow rating, which only **Hilly** reads. See the
-[Mainline deck](StationMaster-Mainline-Deck-v0.4.5.md) reference for what moves a train's entry
+[Mainline deck](mainline-deck.md) reference for what moves a train's entry
 point. On a normal card, a train must check the entire next Subdivision before entering it:

 - an opposing train normally blocks entry;
@@ -193,9 +193,9 @@ cannot carry this column, which is the argument for generating the page rather t
 | --- | --- | --- | :---: |
 | Interlocking | runningTrackStraight | — | **live** |
 | Facing Point Locks | onCard | interlocking in the district | **dormantSolo** |
-| Yard office | secondaryTrackStraight | — | **live** |
-| Small yard | secondaryTrackStraight | — | **live** |
-| Water column | runningTrackStraight | — | **dormantSolo** |
+| Yard Office | secondaryTrackStraight | — | **live** |
+| Small Yard | secondaryTrackStraight | — | **live** |
+| Water Column | runningTrackStraight | — | **dormantSolo** |
 | Overpass | onCard | — | **unbuilt** |
 | Telegraph | runningTrackStraight | — | **live** |
 | Telephone | onCard | telegraph on the same card | **live** |
@@ -68,7 +68,7 @@ read "Both" for both of them, and that was the *other half* of the same mistaken
 House" named those two, §9.3 had to be describing them, so they had to be two-way. Once the Freight
 House is its own card the argument evaporates, and playtesting settled it — "Grocer's Warehouse
 should be receive only, does not ship anything out"; "Refinery: only ships out tanks, does not
-receive anything" (Jesse). `StationMaster-Home-Deck-v0.4.5.md` prints both that way, and the modifier
+receive anything" (Jesse). `home-deck.md` prints both that way, and the modifier
 set agrees: all three Refinery modifiers (Pipelines, Oil Depot, Viscosity Breakers) grant **+1
 outbound**, which would be an odd card set for a facility that receives half the time.
 <!-- TODO v0.5.0: Mine Tipple, Produce Shed and Power Plant above (3/3/4, 2/2/3, 3/3/4) were NOT
@@ -142,7 +142,7 @@ asked a third time:
  rising to 50% at four players. A pre-game interrupt for a rule four games in five never see.

 The docs were the actual defect. `README.md` still listed it among three open rules questions (all
-three closed in v0.5.0) and `StationMaster-Mainline-Deck-v0.4.5.md` still said "the implementation
+three closed in v0.5.0) and `mainline-deck.md` still said "the implementation
 needs a player-selection step to match the card". Both now say settled, and why.

 **Still not implemented**: the Action (10) and Space-use (12) cards, which are genuinely
@@ -555,7 +555,7 @@ Hotel) are what grow them.
 now carries **2** coaches and The Sparrow **3**. The table above is left as `Trains3.pdf` prints it,
 because that is what this section is for — what the design SAYS. What the game plays is
 `src/engine/content.ts`, with the per-card table in
-[`../StationMaster-Home-Deck-v0.4.5.md`](../StationMaster-Home-Deck-v0.4.5.md).
+[`../home-deck.md`](../home-deck.md).

 ### Extras (X13–X22) — ten distinct trains, not four generic ones

@@ -910,7 +910,7 @@ any setting** — being a place an Extra can start is part of what upgrading buy
 - **The Mainline cards were rolled, not dealt.** `buildDivision` drew uniformly from the nine card
  TYPES **with replacement**, so a Division could be dealt two Interchanges or two Tunnels, and
  Plains — printed twice in the deck — carried the same weight as cards printed once.
-  `docs/StationMaster-Mainline-Deck-v0.4.5.md` had already flagged the mismatch as needing
+  `docs/mainline-deck.md` had already flagged the mismatch as needing
  correction; "an Extra may start at the Interchange if one is on the board" is what forced it, since
  that only reads as a rule if the board holds at most one. Now dealt from the printed ten-card deck
  without replacement.
@@ -1,6 +1,6 @@
 {
  "name": "station-master",
-  "version": "0.8.0.15",
+  "version": "0.8.1.0",
  "private": true,
  "type": "module",
  "description": "Station Master — a railroad operations game",
@@ -188,6 +188,56 @@ if (existsSync(imageSrc)) {
  for (const f of readdirSync(imageSrc)) copyFileSync(join(imageSrc, f), join(imageOut, f));
 }

+/**
+ * The player-facing documentation, published beside the game so a tester can reach it from the box.
+ *
+ * COPIED, NOT RE-WRITTEN. The Markdown in `docs/` is the one copy; a hand-written HTML twin would
+ * drift from it on the first edit, which is the whole lesson of TODO #15a and of the 2026-09-20
+ * documentation pass that found four references a month out of date.
+ *
+ * THE WHOLE SET, NOT ONLY THE QUICKSTART. v0.8.0.16 published the guide alone, and the guide's own
+ * §8 "Where to read more" links five further documents by relative path — so every one of them
+ * 404'd on the package (verified on the box: 5 of 6 paths missing). Publishing the guide without
+ * what it points at is the same broken-link failure the test below was written to catch, one hop
+ * further out. The names are kept exactly as the guide writes them, because those links are what
+ * has to resolve.
+ *
+ * SERVED AS PLAIN TEXT for now, which is honest rather than good: tables render as pipes and the
+ * links do not click. Rendering them into styled pages needs a small Markdown converter and is
+ * filed as TODO #109 — this is the version that gets the references in front of testers for this
+ * round rather than leaving them without any.
+ *
+ * PUBLISHED UNDER THEIR OWN NAMES, which since v0.8.0.17 carry no version: the documents are kept
+ * current with every release rather than published as editions, so `docs/rules.md` is served as
+ * `rules.md` and the splash page and the This Game card link it by that name. Four of them were
+ * `StationMaster-<name>-v0.4.5.md` until then — the prototype edition they were first written
+ * against, never the version they described.
+ */
+const GUIDE_DOCS: readonly string[] = [
+  'quickstart.md',
+  'rules.md',
+  'home-deck.md',
+  'mainline-deck.md',
+  'components.md',
+  // Generated by `build:cards` and checked in; a test fails when it disagrees with the code, which
+  // is why it is the one reference that has never drifted. Its `rules/` directory is preserved
+  // because that is the path the Quickstart links it by.
+  'rules/as-built.md',
+];
+
+for (const rel of GUIDE_DOCS) {
+  const src = join(root, 'docs', rel);
+  if (!existsSync(src)) {
+    // Loud rather than silent: a missing document is a broken link on a page already published, and
+    // the build is the only place that can still notice.
+    console.error(`WARNING: docs/${rel} is missing — a published link will 404`);
+    continue;
+  }
+  const out = join(dist, rel);
+  mkdirSync(dirname(out), { recursive: true });
+  copyFileSync(src, out);
+}
+
 // A tiny note for whoever unzips this later and wonders what it needs.
 writeFileSync(
  join(dist, 'README.txt'),
@@ -1825,9 +1825,21 @@ function execute(s: GameState, player: PlayerIndex, i: Intent): GameEvent[] {
    }

    case 'draw.end':
-    case 'freightAgent.end':
      return [{ type: 'phaseEnded', player, phase: 'localOps' }];

+    case 'freightAgent.end':
+      /**
+       * An idle Freight Agent SAYS SO. §6.3 requires no action and choosing to take none is a real
+       * decision — see `freightAgentIdled`. Without the line the log announced the option and then
+       * fell silent, which reads as the game having dropped the turn.
+       */
+      return [
+        ...(turnOf(s, player).freightAgentUsed
+          ? []
+          : [{ type: 'freightAgentIdled', player } as const]),
+        { type: 'phaseEnded', player, phase: 'localOps' },
+      ];
+
    case 'draw.fromHomeOffice': {
      const events: GameEvent[] = [
        {
@@ -2549,8 +2561,28 @@ export function reduce(s: GameState, e: GameEvent): void {
      const f = facilityAt(s, e.player, e.at)!;
      const idx = f.inboundBox.findIndex((c) => c.type === e.stock.type && c.loaded === e.stock.loaded);
      if (idx >= 0) f.inboundBox.splice(idx, 1);
-      // `pooled` — a car back in a yard is back in the common supply, carrying nothing (state.ts).
-      s.yards.classificationYard.push(pooled(e.stock));
+      /**
+       * THE CAR COMES BACK EMPTY, and until v0.8.1.0 it did not.
+       *
+       * The line here said "`pooled` — a car back in a yard is back in the common supply, carrying
+       * nothing", and `pooled` does no such thing: it strips the `origin` stamp and DELIBERATELY
+       * keeps `loaded`, because a train can retire at a Division Point with freight still aboard
+       * (`advance.ts`). So the comment described an intent the call never carried out, and every
+       * car the Freight Agent cleared went to the Classification Yard still loaded.
+       *
+       * That is not cosmetic. `passengersDetrained` needs `type === 'coach' && !loaded` out of the
+       * Division Yard, and §2.2 refills the Division Yard from Classification — so a coach cleared
+       * from a red box came back as stock that could never unload another passenger. Measured over
+       * five three-Day solitaire games before the fix: 18 loaded coaches in the Classification Yard
+       * against 6 empty, and 9 loaded in the Division Yard against 33 empty.
+       *
+       * The red box is where the journey ENDS for what is in it — the Revenue was paid when the
+       * load arrived or the passengers detrained (`view.ts`, `freightAgent.clearInbound`). Clearing
+       * it sends the passengers out of the station, or the delivered load into the industry, and
+       * returns the CAR to the common supply. A car in a yard carrying a load nobody can unload is
+       * not supply.
+       */
+      s.yards.classificationYard.push({ ...pooled(e.stock), loaded: false });
      turnOf(s, e.player).freightAgentUsed = true;
      break;
    }
@@ -289,7 +289,7 @@ export const INDUSTRY_PROFILES: readonly IndustryProfile[] = [
   * The card set says the same thing on its own. All three Refinery modifiers — Pipelines, Oil
   * Depot, Viscosity Breakers — grant `+1 outbound`; a two-way Refinery would be the one industry in
   * the game with no way to raise the direction it is supposed to use half its capacity on.
-   * `StationMaster-Home-Deck-v0.4.5.md` prints it "Outbound, 1 out / 0 in".
+   * `home-deck.md` prints it "Outbound, 1 out / 0 in".
   */
  { kind: 'refinery', name: 'Refinery', carTypes: ['tank'], flow: 'outbound', baseOut: 1, baseIn: 0, baseLoaders: 1, lockouts: ['powerPlant'], copies: 1 },
  { kind: 'powerPlant', name: 'Power Plant', carTypes: ['hopper', 'tank'], flow: 'inbound', baseOut: 0, baseIn: 1, baseLoaders: 1, lockouts: ['mineTipple', 'refinery'], copies: 2 },
@@ -297,7 +297,7 @@ export const INDUSTRY_PROFILES: readonly IndustryProfile[] = [
  /**
   * INBOUND ONLY — the mirror of the Refinery above, and the same correction. Reported from
   * playtesting and confirmed by Jesse (v0.4.9e): "Grocer's Warehouse should be receive only, does
-   * not ship anything out". `StationMaster-Home-Deck-v0.4.5.md` prints it "Inbound, 0 out / 1 in".
+   * not ship anything out". `home-deck.md` prints it "Inbound, 0 out / 1 in".
   *
   * THE ICE HOUSE IS THEREFORE A DEAD CARD BESIDE A GROCER'S, and that is the design, not an
   * oversight: `usableGrant` drops a Modifier's grant on a direction its host cannot use, and the
@@ -589,7 +589,20 @@ export type MainlineProfile = {
  speedStarts?: { fast: number; slow: number };
  /** Double Track: "Trains may pass". */
  trainsMayPass: boolean;
-  /** Interchange: "Sort cars in new order". */
+  /**
+   * Interchange only. The card prints "Sort cars in new order" — **and that is not what this flag
+   * does**, which is why it is worth spelling out where the field is declared.
+   *
+   * The printed sorting has never been implemented: nothing reads this to permit a sort, and a
+   * consist is re-ordered at a Small Yard in a district (`switch.sortConsist`). What this actually
+   * marks is the one Mainline card with a Yard Limit, and therefore the one an Extra may be made up
+   * and started on (`apply.ts` § resolveExtraStart, `legal.ts`).
+   *
+   * Named for the printed text, and kept that way deliberately — renaming it would lose the link to
+   * the card face — but the name has already misled once: `mainlineDescription` grew a sentence
+   * telling players cars could be sorted here, which reached the board and the generated card
+   * reference before it was caught on 2026-09-20.
+   */
  sortsCars: boolean;
  /** Named entry points printed on the card; some are unlocked by modifier cards. */
  entryPoints: readonly string[];
@@ -655,7 +668,7 @@ export const MAINLINE_PROFILES: readonly MainlineProfile[] = [
 * used it as one: `buildDivision` drew uniformly from those types with replacement, which made two
 * Interchanges (or two Trestles, or two Tunnels) an ordinary outcome and gave Plains the same weight
 * as everything else although the deck prints two of it. `Mainline Cards.pdf` is the
- * inventory, transcribed in `docs/StationMaster-Mainline-Deck-v0.4.5.md`, which had already flagged
+ * inventory, transcribed in `docs/mainline-deck.md`, which had already flagged
 * the mismatch as needing correction.
 *
 * It matters more than card flavour now that an Extra may start at an Interchange (§7): "if an
@@ -995,14 +1008,25 @@ export function enhancementRule(key: string): EnhancementRule | null {
  return ENHANCEMENT_RULES.find((r) => r.key === key) ?? null;
 }

+/**
+ * THE NAMES HERE ARE WHAT A PLAYER READS, since v0.8.0.17 — `cardName` takes them in preference to
+ * `prettyKey`'s guess, which cannot recover an acronym (it rendered `absSignals` as "Abs Signals"
+ * beside a tooltip saying ABS).
+ *
+ * So their CASE matters, and three were transcribed in sentence case while the rules, the panels
+ * and every other line of this repository treat them as proper terms: "Yard Office" outnumbered
+ * "Yard office" 36 to 2, "Small Yard" 36 to 2, "Water Column" 10 to 2. Corrected here rather than
+ * special-cased in `cardName`, because this table is the source and a lookup that second-guesses
+ * its own source is the drift it exists to prevent.
+ */
 export const ENHANCEMENT_CARDS: readonly SimpleCard[] = [
  { key: 'interlocking', name: 'Interlocking', copies: 1, placement: 'any Running Track Straight', effect: 'May stop an inbound train on the Limit Track.' },
  // Not in sheet 5 — dealt 0 copies (Jesse, 2026-08-26). It answers Derail, which is itself an
  // Event held out until built, so at zero it defends against nothing that can be dealt anyway.
  { key: 'facingPointLocks', name: 'Facing Point Locks', copies: 0, placement: 'adjacent to Interlocking', effect: 'Must have Interlocking. Prevents Derail being played on you.', answers: 'Derail' },
-  { key: 'yardOffice', name: 'Yard office', copies: 1, placement: 'any Secondary Track Straight', effect: 'An inbound train with no coaches that can reach the yard office in one move may arrive there instead of the Train Order Office.' },
-  { key: 'smallYard', name: 'Small yard', copies: 1, placement: 'any Secondary Track Straight', effect: 'A train that spends one move in the yard may sort itself into ANY order, including cars ahead of the engine.' },
-  { key: 'waterColumn', name: 'Water column', copies: 1, placement: 'any Running Track Straight', effect: 'Lets you remove any Watertower in your district.', answers: 'Watertower' },
+  { key: 'yardOffice', name: 'Yard Office', copies: 1, placement: 'any Secondary Track Straight', effect: 'An inbound train with no coaches that can reach the yard office in one move may arrive there instead of the Train Order Office.' },
+  { key: 'smallYard', name: 'Small Yard', copies: 1, placement: 'any Secondary Track Straight', effect: 'A train that spends one move in the yard may sort itself into ANY order, including cars ahead of the engine.' },
+  { key: 'waterColumn', name: 'Water Column', copies: 1, placement: 'any Running Track Straight', effect: 'Lets you remove any Watertower in your district.', answers: 'Watertower' },
  { key: 'overpass', name: 'Overpass', copies: 1, placement: 'any Railroad Crossing', effect: 'Removes the restrictions of a played Railroad Crossing.', answers: 'Railroad crossing' },
  /**
   * THE DISPATCHING LADDER IS OUT OF THE DECK, at 0 copies rather than deleted — the treatment
@@ -115,6 +115,17 @@ export type GameEvent =
      movesAllowed: number;
      lastMove?: { trayId: TrayId; to: GridCoord };
    }
+  /**
+   * THE FREIGHT AGENT WAS CHOSEN AND DID NOTHING, which is a decision rather than an absence.
+   *
+   * §6.3 requires no action, and the bot deliberately takes that route — unjamming a healthy box
+   * destroys a load that cost a whole Local Operations action to stock, so an idle Stage is
+   * strictly better. Reported from a table on Day 1 Stage 3 of v0.8.0.16: the log announced
+   * FREIGHT AGENT work and then said nothing at all, so the turn read as a bug.
+   *
+   * Reduces to nothing, like `switchingEnded` above: it reports a choice the state already holds.
+   */
+  | { type: 'freightAgentIdled'; player: PlayerIndex }
  | { type: 'cardDrawn'; player: PlayerIndex; source: 'homeOffice' | 'department'; slot?: number; cardId: CardId }
  /**
   * §6.2 — the Home Office deck ran out, so the Salvage Yard and all three Department decks were
@@ -230,7 +230,7 @@ function buildPassengerFacility(tier: Parameters<typeof officeProfile>[0]): NonN
 * `MAINLINE_PROFILES` is a list of card TYPES and this drew from it uniformly WITH replacement, so
 * a Division could be handed two Interchanges or two Tunnels, and Plains — printed twice in the
 * deck — carried the same weight as cards printed once. `MAINLINE_DECK` is the printed inventory
- * (`docs/StationMaster-Mainline-Deck-v0.4.5.md`, which flagged this as needing correction), and the
+ * (`docs/mainline-deck.md`, which flagged this as needing correction), and the
 * deal is now a deal: take cards out of it and do not put them back.
 *
 * The Extra-start rules are what forced the issue. "An Extra may start at the Interchange if one is
@@ -85,6 +85,14 @@ const MIME: Record<string, string> = {
  '.json': 'application/json; charset=utf-8',
  '.png': 'image/png',
  '.svg': 'image/svg+xml',
+  /**
+   * The Quickstart guide, published by `build-web.ts` as `quickstart.md`.
+   *
+   * text/plain ON PURPOSE. The fallback below is `application/octet-stream`, which makes a browser
+   * DOWNLOAD the file instead of showing it — so without this line the splash page's "read the
+   * guide" link hands a tester a file to save rather than a page to read.
+   */
+  '.md': 'text/plain; charset=utf-8',
 };

 const HEARTBEAT_MS = 20_000;
@@ -71,6 +71,22 @@ export function divisionSvg(nodes: DivisionView[], roster?: DivisionRoster | nul
   * Self-contained on purpose: the replay embeds this by `toString()`, so it may not reach for
   * anything outside its own body.
   */
+  /**
+   * The Mainline modifiers, as the three letters drawn on a card.
+   *
+   * DECLARED INSIDE THIS FUNCTION, like every constant around it, because `sim/replay.ts` emits
+   * `divisionSvg.toString()` into the replay page — a module-scope const it closed over would be
+   * undefined there, and the page threw exactly that way before this was moved.
+   *
+   * Keyed by the name the VIEW builds (`prettyKey`'d in `view.ts`), because that is what arrives
+   * here — not the engine's key, which this file never sees. A modifier with no tag draws nothing
+   * rather than a raw key: `flatMap` over a missing entry yields none, so a card added upstream is
+   * silently unmarked instead of printing "rotaryDumps" at a player.
+   *
+   * ABS Signals is deliberately absent: it has a signal mast of its own.
+   */
+  const MOD_TAGS: Record<string, string> = { Brakeman: 'BRK', Airbrakes: 'AIR', Helpers: 'HLP' };
+
  const CW = { dp: 118, ml: 152, run: 78 };
  /**
   * TALL ENOUGH FOR TWO REGISTERS OF CHIPS, on every cell so the rail runs level across the row.
@@ -144,6 +160,10 @@ export function divisionSvg(nodes: DivisionView[], roster?: DivisionRoster | nul
    tip: string;
    /** Set on a card that has just changed under the players' feet — drawn with a brief pulse. */
    flash?: boolean;
+    /** ABS Signals standing on this Mainline card — drawn as a signal mast, not only described. */
+    abs?: boolean;
+    /** The Mainline modifiers on this card, as short tags — see `MOD_TAGS`. */
+    mods?: string[];
    /** Which SEAT's district this cell belongs to, or null for Mainline and Division Points. */
    seat: number | null;
    /** Set on an Office cell when a roster was supplied: whose district this is. */
@@ -291,6 +311,10 @@ export function divisionSvg(nodes: DivisionView[], roster?: DivisionRoster | nul
          // words with no gameplay attached — reported exactly that way.
          (n.what ? `\n\n${n.what}` : ''),
      seat: null,
+      // Attachments are drawn, not only described: `modifiers` carries ABS Signals alongside the
+      // Mainline modifiers, and the name is the one the view already built for the tooltip.
+      abs: !dp && n.modifiers.includes('ABS Signals'),
+      mods: dp ? [] : n.modifiers.flatMap((m) => MOD_TAGS[m] ?? []),
      // A Division Point is one region — the queue trains enter and leave the Division through.
      regions: dp ? 1 : (n.regions ?? 0),
      gradeUp: dp ? null : (n.gradeUp ?? null),
@@ -416,6 +440,57 @@ export function divisionSvg(nodes: DivisionView[], roster?: DivisionRoster | nul
      out += `</g>`;
    }

+    /**
+     * ABS SIGNALS, DRAWN RATHER THAN ONLY SAID — reported from a table on Day 1 Stage 1 of
+     * v0.8.0.16: "when played on the trestle, there was no on-the-card indication; it's only when
+     * you look at the tooltip for trestle that you see that ABS exists."
+     *
+     * Exactly the complaint the grade wedge below answers, and the card it protects is the one a
+     * player is deciding whether to run a second train onto. A signal is the literal object — ABS
+     * is Automatic Block Signals — so the mark is a mast with a lit lamp rather than a badge
+     * reading "ABS", and it needs no room for text, which is what lets it sit clear of a name as
+     * long as "Uncontrolled Siding" on a 152px card.
+     *
+     * TOP-RIGHT, above the rail and clear of the chip registers: the flag owns the cell ends at
+     * rail height and the grade wedge owns the bottom-right, so this is the corner left.
+     */
+    if (c.abs) {
+      const mx = c.x + c.w - 10;
+      const top = c.y + 5;
+      out += `<g class="bs-abs">`;
+      out += `<line class="bs-abs-mast" x1="${mx}" y1="${top}" x2="${mx}" y2="${top + 15}"/>`;
+      // Two lamps, the upper one lit: a signal showing an aspect, not a bare post.
+      out += `<circle class="bs-abs-lit" cx="${mx}" cy="${top + 3}" r="2.6"/>`;
+      out += `<circle class="bs-abs-dark" cx="${mx}" cy="${top + 9}" r="2.6"/>`;
+      out += `</g>`;
+    }
+
+    /**
+     * THE MAINLINE MODIFIERS, DRAWN RATHER THAN ONLY SAID (Jesse, 2026-09-21) — the other half of
+     * the ABS report above: "Brakeman, Airbrakes, Helpers and Realignment should also be drawn on
+     * the card, not just the tooltip."
+     *
+     * ONLY THREE OF THOSE FOUR CAN EVER BE HERE. Realignment does not sit on a card — `reduce`
+     * takes the `became` branch and CHANGES `node.card`, so a realigned Trestle simply is an
+     * Uncontrolled Siding afterwards and the card face already says so. The other three are
+     * `gradeOnly`, so in practice this is a Heavy Grade's row.
+     *
+     * TAGS, NOT NAMES, and only because the measurements leave no choice: "Brakeman · Airbrakes ·
+     * Helpers" is 30 characters where about eleven fit beside a card name on a 152px cell. Three
+     * letters is the most that fits while still mapping to one card each, and the tooltip — which
+     * has always named them in full — is what expands it. The mark says THAT there is one, which
+     * is the half that was missing.
+     *
+     * Right-aligned on the name row: the only band on the cell that is clear, with the chip
+     * registers starting at `CHIP_Y` below and the capacity line and grade wedge at the foot.
+     * Shifted left of the signal when a card carries both.
+     */
+    if (c.mods && c.mods.length > 0) {
+      out +=
+        `<text class="bs-mod" text-anchor="end" x="${c.x + c.w - (c.abs ? 20 : 7)}" y="${c.y + 14}">` +
+        `${esc(c.mods.join('·'))}</text>`;
+    }
+
    /**
     * WHICH WAY A HEAVY GRADE CLIMBS, drawn rather than only said.
     *
@@ -1392,6 +1467,12 @@ export const BOARD_CSS = `
 .bs-grade{fill:#e08060;font:10px ui-monospace,monospace}
 .bs-mod{font:10px ui-monospace,monospace}
 text.bs-mod{fill:#c8a04a}
+/* A signal standing on a Mainline card carrying ABS Signals. Green is otherwise unused on the
+   Division row apart from the Division Point's dashed border, so a lit lamp does not compete with
+   the amber "it is happening here" or the red flag for meaning. */
+.bs-abs-mast{stroke:#9aa3b0;stroke-width:1.6}
+.bs-abs-lit{fill:#4fae6a;stroke:#2c6b40;stroke-width:0.8}
+.bs-abs-dark{fill:#2a3038;stroke:#59626f;stroke-width:0.8}
 .bs-enh{fill:#7fb0e6;font:9px ui-monospace,monospace}
 .bs-enh-spent{fill:#5b6b7d;text-decoration:line-through}
 .bs-rowlab{fill:#5f6b7a;font:600 9px ui-monospace,monospace;letter-spacing:.1em}
@@ -74,7 +74,16 @@ export function carsLabel(cars: RollingStock[]): string {
  return cars.map(carLabel).join(', ');
 }

-const at = (c: GridCoord): string => `(${c.row},${c.col})`;
+/**
+ * X,Y — EAST/WEST THEN NORTH/SOUTH, exactly as `view.ts` writes it, and NOT the internal row/col
+ * storage order.
+ *
+ * These two disagreed until 2026-09-21: the action menu said "(1,-1)" and the log said "(-1,1)" for
+ * the same square, side by side on the same screen. `view.ts` carried the comment explaining why
+ * the display order is X,Y; this one had no comment at all and was simply the storage order
+ * reaching the page. Jesse's call — the log and the action menu spell a square the same way.
+ */
+const at = (c: GridCoord): string => `(${c.col},${c.row})`;

 const BOX_NAMES = ['MEN', 'AT', 'WORK'] as const;
 const boxName = (i: number): string => BOX_NAMES[i] ?? `box ${i}`;
@@ -135,6 +144,17 @@ export type NarrateContext = {
   * industry is the whole point of the move, so it is what the line should say.
   */
  facilityAt?: (player: PlayerIndex, at: GridCoord) => string | null;
+  /**
+   * Names the Mainline card at a Division node — "Plains", "Heavy Grade" — or null if it cannot.
+   *
+   * A NODE INDEX IS NOT A NAME. The log said "ENHANCEMENT built: abs Signals on Mainline card 7",
+   * which asks the reader to remember what card 7 was (Jesse, 2026-09-22, playing ABS Signals).
+   * The SLOT still matters and is kept beside it, because a division can hold two Plains and the
+   * number is the only thing that tells them apart.
+   */
+  mainlineAt?: (node: number) => string | null;
+  /** The printed name of an Enhancement or Mainline modifier, by its content key. */
+  enhancementName?: (key: string) => string | null;
 };

 export function narrate(e: GameEvent, ctx: NarrateContext = {}): Narration {
@@ -143,6 +163,12 @@ export function narrate(e: GameEvent, ctx: NarrateContext = {}): Narration {
  // The industry on a square when there is one, and the coordinate when there is not — a crew works
  // plain track too, and "at nowhere" would be worse than the notation.
  const place = (player: PlayerIndex, c: GridCoord): string => ctx.facilityAt?.(player, c) ?? at(c);
+  /** "Mainline card 7 (Plains)" — the slot, which disambiguates, plus the card, which informs. */
+  const mainlineCard = (node: number | undefined): string => {
+    if (node === undefined) return 'the Mainline';
+    const name = ctx.mainlineAt?.(node) ?? null;
+    return `Mainline card ${node}${name ? ` (${name})` : ''}`;
+  };

  switch (e.type) {
    // -- clock
@@ -188,7 +214,15 @@ export function narrate(e: GameEvent, ctx: NarrateContext = {}): Narration {
            ? 'Chose to SWITCH — six Moves to shunt cars around the yard. Watch the crew chip on the grid: it carries its consist with it, and cars it passes over are coupled automatically.'
            : e.option === 'draw'
              ? 'Chose to DRAW a card'
-              : 'Chose FREIGHT AGENT work — one car moved to or from a facility',
+              /**
+               * SAYS WHAT MAY BE DONE, NOT WHAT WAS. This read "one car moved to or from a
+               * facility" — an assertion — and §6.3 requires no action at all, so when the Freight
+               * Agent went idle the log claimed a car had moved and then fell silent about which.
+               * Reported from a table on Day 1 Stage 3 of v0.8.0.16. The work itself is narrated by
+               * `stockToOutbound`, `inboundCleared` and `facilityUnjammed`, each naming the car and
+               * the industry; an idle Agent is narrated by `freightAgentIdled`.
+               */
+              : 'Chose FREIGHT AGENT work — may stock a green Outbound box, clear a red Inbound one, or free a jam',
      };
    case 'trayMoved':
      // `via` rides on the event only when there was another legal route to the same square
@@ -304,7 +338,10 @@ export function narrate(e: GameEvent, ctx: NarrateContext = {}): Narration {
        tone: 'plain',
        text: e.became
          ? `Realignment: Mainline card ${e.node}, ${kindName(e.from)}, converted to ${kindName(e.became)}`
-          : `Played ${e.key} on Mainline card ${e.node}${e.from ? ` (${kindName(e.from)})` : ''}`,
+          // Same two faults as `enhancementPlaced` above, in the neighbouring case: a raw key for a
+          // name, and a slot with nothing to say which card it is when `from` is absent.
+          : `Played ${ctx.enhancementName?.(e.key) ?? e.key} on Mainline card ${e.node}` +
+            ` (${e.from ? kindName(e.from) : (ctx.mainlineAt?.(e.node) ?? 'that card')})`,
      };
    }
    case 'redFlagSpent':
@@ -354,11 +391,17 @@ export function narrate(e: GameEvent, ctx: NarrateContext = {}): Narration {
        text: `Extra X${e.trainNumber} played — it is NOT scheduled; it runs once as soon as a Crew Tray frees up, then its card is gone`,
      };
    case 'enhancementPlaced': {
-      const name = e.key.replace(/([A-Z])/g, ' $1');
+      /**
+       * THE CARD'S OWN NAME, not a key with spaces wedged into it. This was
+       * `e.key.replace(/([A-Z])/g, ' $1')`, which rendered `absSignals` as "abs Signals" —
+       * uncapitalised, and bypassing the authored names `cardName` started honouring in v0.8.0.17
+       * for exactly this reason: an acronym is not recoverable by splitting on capitals.
+       */
+      const name = ctx.enhancementName?.(e.key) ?? e.key.replace(/([A-Z])/g, ' $1');
      // Out on the Mainline is not a square in anyone's district, so it is named rather than
      // given a coordinate the Office Area does not have.
      if (e.at === undefined) {
-        return { tone: 'good', text: `ENHANCEMENT built: ${name} on Mainline card ${e.node}` };
+        return { tone: 'good', text: `ENHANCEMENT built: ${name} on ${mainlineCard(e.node)}` };
      }
      return { tone: 'good', where: e.at, text: `ENHANCEMENT built: ${name} at ${at(e.at)}` };
    }
@@ -466,23 +509,39 @@ export function narrate(e: GameEvent, ctx: NarrateContext = {}): Narration {
      };

    // -- freight agent
+    /**
+     * THE INDUSTRY, NOT THE COORDINATE — `place` over `at`, for the reason its own comment gives:
+     * "(-1,1)" is the grid's notation and means nothing at a table where people are looking at
+     * cards. The switching lines were moved to it and these three were missed, so a Freight Agent
+     * turn was the one place the log still spoke in coordinates. Asked directly from a table on
+     * Day 1 Stage 3 of v0.8.0.16: "can we tell what car, what facility and whether it was to or
+     * from." The car and the direction were already here; the facility was not.
+     */
    case 'stockToOutbound':
      return {
        tone: 'plain',
        where: e.at,
-        text: `Freight Agent put a ${carLabel(e.stock)} into the green Outbound box at ${at(e.at)}`,
+        text: `Freight Agent loaded a ${carLabel(e.stock)} INTO the green Outbound box at ${place(e.player, e.at)}`,
      };
    case 'inboundCleared':
      return {
        tone: 'plain',
        where: e.at,
-        text: `Freight Agent cleared a ${carLabel(e.stock)} from the red Inbound box at ${at(e.at)}`,
+        text: `Freight Agent cleared a ${carLabel(e.stock)} OUT of the red Inbound box at ${place(e.player, e.at)}`,
      };
    case 'facilityUnjammed':
      return {
        tone: 'bad',
+        text: `UNJAMMED ${place(e.player, e.at)} — pulled a ${carLabel(e.stock)} out of ${e.from} to free the facility`,
        where: e.at,
-        text: `UNJAMMED ${at(e.at)} — pulled a ${carLabel(e.stock)} out of ${e.from} to free the facility`,
+      };
+    case 'freightAgentIdled':
+      return {
+        tone: 'quiet',
+        text:
+          'Freight Agent found nothing worth doing — no green box could be stocked, no red box needed ' +
+          'clearing, and no load was jammed. §6.3 requires no action, and unjamming a healthy box ' +
+          'would destroy a load that cost a whole action to stock.',
      };

    // -- trains
@@ -953,9 +953,22 @@ export function describeIntent(s: GameState, i: Intent): string {
        ? areaOf(s, actor).grid.get(`${i.placement.row},${i.placement.col}`)
        : undefined;
      const upgrade = over?.geometry.kind === 'track';
+      /**
+       * NAME THE MAINLINE CARD, for exactly the reason the rotation is named above — reported from
+       * a table on Day 1 Stage 1 of v0.8.0.16 and the THIRD time this trap has been sprung.
+       *
+       * ABS Signals is played on a Division NODE rather than a grid square, so `i.placement` is
+       * absent and every one of its placements described itself as plain "play ABS Signals". The
+       * action list drops duplicate labels, so all but the lowest-index Mainline card were discarded
+       * before the menu saw them: the tooltip promised "any Mainline card" and the board offered
+       * one. Naming the card is what makes the choice both legible and survivable.
+       */
+      const onNode = i.node === undefined ? undefined : s.division.nodes[i.node];
+      const mainline =
+        onNode?.kind === 'mainline' ? ` on the ${mainlineProfile(onNode.card).name}, out on the Mainline` : '';
      return (
        `${upgrade ? 'upgrade to' : 'play'} ${cardName(s, i.cardId)}` +
-        `${i.placement ? ` at ${at(i.placement)}` : ''}${turn}`
+        `${i.placement ? ` at ${at(i.placement)}` : ''}${mainline}${turn}`
      );
    }
    case 'card.discard': {
@@ -1137,8 +1150,20 @@ export function describeIntent(s: GameState, i: Intent): string {
          ? 'the Revenue was paid when the passengers detrained'
          : 'the Revenue was paid when the load reached the box';
      const frees = f?.kind === 'passenger' ? 'more passengers can detrain here' : 'another car can be unloaded here';
+      /**
+       * SAY WHAT HAPPENS, NOT WHAT MOVES (Jesse, 2026-09-22). This read "send the loaded coach in
+       * the red Inbound box to the Classification Yard", which is true of the counters and
+       * nonsense as a sentence — it sounds like shipping a coachful of people off to a siding.
+       * What the red box holds is a journey that has ENDED: the passengers are at the station, or
+       * the load is at the industry, and both were paid for on arrival. Clearing it sends them on
+       * their way and returns the CAR to the common supply, empty.
+       */
+      const what =
+        f?.kind === 'passenger'
+          ? `let the passengers out of the station — the ${car ? carLabel(car) : 'coach'} in the red Inbound box at ${at(i.at)} goes back to the Classification Yard empty`
+          : `send the delivered load into the industry — the ${car ? carLabel(car) : 'car'} in the red Inbound box at ${at(i.at)} goes back to the Classification Yard empty`;
      return (
-        `send the ${car ? carLabel(car) : 'car'} in the red Inbound box at ${at(i.at)} to the Classification Yard` +
+        `${what}` +
        ` — pays nothing (${paid}); it frees${full ? ' the last' : ' a'} slot so ${frees}`
      );
    }
@@ -1903,10 +1928,46 @@ export function cardName(s: GameState, id: string): string {
    case 'mainlineModifier':
    case 'maneuver':
    case 'action':
-      return prettyKey(k.key);
+      /**
+       * THE PRINTED NAME WINS over `prettyKey`'s guess, on the same reasoning as the facility and
+       * modifier lookups above: the tables are for names `prettyKey` cannot derive.
+       *
+       * It guessed wrong more often than the fallback comment implies. `absSignals` came out as
+       * "Abs Signals" on the button while the rules, the tooltip and the card face all say **ABS**
+       * Signals — an acronym no key-splitter can recover — and `brokenCoupler` and `beanHouse`
+       * were title-cased past their printed "Broken coupler" and "Bean house".
+       */
+      return SIMPLE_CARD_NAMES.get(k.key) ?? prettyKey(k.key);
  }
 }

+/**
+ * Every authored card name, keyed as the card kinds key themselves.
+ *
+ * Built from the content tables rather than restated, so a card renamed there is renamed here and
+ * the two cannot drift — which is the whole argument of TODO #15a, applied to names.
+ */
+/**
+ * The printed name of a card that has one, by its content key, or null.
+ *
+ * Exported so the LOG can reach the same table the buttons do. `narrate.ts` rendered its own name
+ * from the key (`absSignals` → "abs Signals") and so disagreed with the action list one line above
+ * it; it takes this through its `enhancementName` resolver now.
+ */
+export function simpleCardName(key: string): string | null {
+  return SIMPLE_CARD_NAMES.get(key) ?? null;
+}
+
+const SIMPLE_CARD_NAMES: ReadonlyMap<string, string> = new Map(
+  [
+    ...SPACE_USE_CARDS,
+    ...ENHANCEMENT_CARDS,
+    ...MAINLINE_MODIFIER_CARDS,
+    ...MANEUVER_CARDS,
+    ...ACTION_CARDS,
+  ].map((c) => [c.key, c.name]),
+);
+
 /**
 * What a card actually DOES, in one line.
 *
@@ -37,16 +37,7 @@ import { collectStep, newCollector } from '../sim/display-step.ts';
 import type { DisplayCollector } from '../sim/display-step.ts';
 // Import from the view module, NOT replay.ts — replay.ts writes files and reads process.argv,
 // which would pull node:fs into a browser bundle.
-import {
-  cardDescription,
-  cardName,
-  currentActorOfState,
-  describeIntent,
-  geometryLabel,
-  snapshot,
-  trainName,
-  variantLabel,
-} from '../sim/view.ts';
+import { cardDescription, cardName, currentActorOfState, describeIntent, geometryLabel, simpleCardName, snapshot, trainName, variantLabel } from '../sim/view.ts';
 import {
  DEFAULT_DAYS,
  DEFAULT_HOUSE_RULES,
@@ -1354,6 +1345,19 @@ function record(game: Game, events: GameEvent[], actor: PlayerIndex | null = nul
      cardName: (id) => cardName(game.state, id),
      trainName: (id) => trainName(game.state, id),
      facilityAt: (player, at) => facilityOn(game, player, at),
+      /**
+       * Which Mainline card a Division node holds. The log used to print the node INDEX alone —
+       * "on Mainline card 7" — leaving the reader to remember what card 7 was (Jesse, 2026-09-22).
+       * The index is kept beside the name because a division can deal two Plains, and then the slot
+       * is the only thing that distinguishes them.
+       */
+      mainlineAt: (node) => {
+        const n = game.state.division.nodes[node];
+        return n?.kind === 'mainline' ? mainlineProfile(n.card).name : null;
+      },
+      // The printed name, for the same reason `cardName` stopped trusting `prettyKey` in v0.8.0.17:
+      // no key-splitter recovers "ABS Signals" from `absSignals`.
+      enhancementName: (key) => simpleCardName(key),
      // Whose district a train reached is not the actor — the Mainline Phase has none — so the
      // narration resolves the name itself rather than being prefixed with one by the code below.
      // NO NUMBER IN THE FALLBACK. This is a PLAYER index, and a player is not a seat — seats rotate
@@ -27,6 +27,11 @@ h1{font-size:32px;margin:0 0 2px;letter-spacing:.02em}
 a.door:hover{border-color:#4d6fa8;background:#1f2733;transform:translateY(-1px)}
 .door h2{font-size:17px;margin:0 0 5px;color:#9fb6d8}
 .door p{margin:0;color:var(--dim);font-size:13px;line-height:1.5}
+/* Not a fourth door: reading the guide is not a way to play, and giving it equal weight in the
+   grid would say it is. A line under the doors, where somebody who does not know what to click
+   will already be looking. */
+.newhere{margin:16px 2px 0;color:var(--dim);font-size:13px;line-height:1.55}
+.newhere a{color:#9fb6d8}
 .door .go{display:inline-block;margin-top:11px;font-size:12px;color:#5aa9e6}
 .door.disabled .go{color:var(--dim)}
 a.door.disabled{pointer-events:none}
@@ -96,6 +101,11 @@ footer{margin-top:26px;color:var(--dim);font-size:11px;display:flex;gap:18px;fle
    </a>
  </div>

+  <p class="newhere">New to Station Master?
+    <a href="./quickstart.md">Read the Quickstart guide</a> &mdash; what the game is, how you win,
+    how a Stage runs, what is on the screen, and a first twenty minutes. About twenty minutes to
+    read, and it will save you an hour of guessing.</p>
+
  <div class="rule"></div>

  <footer>
@@ -131,6 +131,22 @@ export function runLobby(handlers: LobbyHandlers, resume?: { token: string; game
  // stub only part of the DOM. Bail rather than throwing through the module's caller.
  if (!has('lobby') || !has('lb-choice-section')) return;
  $('lobby').hidden = false;
+  /**
+   * THE TWO DOORS COME BACK EVERY TIME THIS SCREEN DOES.
+   *
+   * `enterSeating` hides this section, and until v0.8.1.0 the only things that put it back were the
+   * lobby's OWN leave paths. Leaving a running game is a different route — `main.ts`'s Leave game
+   * button calls `runLobby` again — so the section stayed hidden from the previous seating and the
+   * screen came back holding nothing but "Games you are in". Reported 2026-09-22: "I see the games
+   * that I'm in, and I have the ability to rejoin or forget. But what if I want to join a different
+   * game or start a new one? How do I get back there?" There was no way: the doors were on the
+   * page, display:none, with no control that could reveal them.
+   *
+   * Reset here rather than in the leave handler, because this function is the one thing every route
+   * onto the lobby goes through — which is exactly why the two that did it themselves missed a third.
+   */
+  $('lb-choice-section').hidden = false;
+  if (has('lb-seating-section')) $('lb-seating-section').hidden = true;

  const form = settingsForm('lb-');
  let source: EventSource | null = null;
@@ -652,9 +652,48 @@ function renderGameCard(f: Frame): void {
  const code = gameCode === '' ? '' : `<dt>Game code</dt><dd>${esc(gameCode)}</dd>`;
  $('gamecardbody').innerHTML =
    `<dl>${who}${code}<dt>Type</dt><dd>${esc(gameTypeLabel(type, f.mode))}</dd></dl>` +
-    rulesListHtml(config, players, f.days);
+    rulesListHtml(config, players, f.days) +
+    GUIDE_HTML;
 }

+/**
+ * THE DOCUMENTATION, REACHABLE FROM INSIDE A GAME — asked for directly from a table, 2026-09-21:
+ * "how can we link the documentation so it can be reached from the gameplay, whether someone is
+ * playing solitaire or multiplayer?"
+ *
+ * IT NEEDS NO MODE AWARENESS, which is the whole reason this is three lines rather than a feature.
+ * Solitaire and multiplayer are the same page on the same origin — `play.html?solitaire` and
+ * `play.html?lobby` — and the build publishes the documents beside it, so one relative link
+ * resolves identically in both, on the public site and on a StartOS box alike.
+ *
+ * IN THIS CARD RATHER THAN THE HEADER (Jesse's call, 2026-09-21), for the reason he gave the top
+ * line in the first place on 2026-08-30: it is "not something that they're likely to need all the
+ * time", and the header is the line that must not wrap. The card is already where reference lives
+ * — the seed, the seat, the house rules — and it holds the whole set rather than one door.
+ *
+ * NEW TAB, every one of them: a player reading the rules mid-turn must not lose the game behind
+ * them. `rel="noopener"` because `target="_blank"` without it hands the opened page a handle back.
+ *
+ * Built once at module scope — it never varies, and rebuilding it on every frame would be work
+ * nobody sees, the same reasoning the folded body above already follows.
+ */
+const GUIDE_DOCS: readonly { href: string; label: string; what: string }[] = [
+  { href: './quickstart.md', label: 'Quickstart', what: 'What the game is and a first twenty minutes — for anyone who has not played.' },
+  { href: './rules.md', label: 'Rules', what: 'The rules in full, with the FAQ.' },
+  { href: './rules/as-built.md', label: 'Every card', what: 'Generated from the code, so it cannot drift from what the game actually does.' },
+  { href: './home-deck.md', label: 'Home deck', what: 'How the Home Office deck is dealt and played.' },
+  { href: './mainline-deck.md', label: 'Mainline deck', what: 'The Mainline cards and what each does to a train.' },
+  { href: './components.md', label: 'Components', what: 'Rolling stock, yards, trays, the Fedora.' },
+];
+
+const GUIDE_HTML =
+  `<div class="guide"><h4>Guide</h4><p>` +
+  GUIDE_DOCS.map(
+    (d) =>
+      `<a href="${esc(d.href)}" target="_blank" rel="noopener" data-tip="${esc(d.what)}">${esc(d.label)}</a>`,
+  ).join(' · ') +
+  `</p></div>`;
+
 /**
 * THE COLLISION COUNTS, WHICH ARE A LIVE SCORE (TODO #28, Jesse's call 2026-08-30).
 *
@@ -268,6 +268,13 @@ button.ghost:hover{border-color:#4d6fa8;color:var(--fg)}
 #gamecardbody dd{margin:0;font-size:12px;color:#cfd6e0}
 #gamecardbody dd.changed{color:#f0b64a}
 #gamecardbody h4{margin:8px 0 0;font-size:11px;text-transform:uppercase;letter-spacing:.06em;color:#8b94a3}
+/* THE GUIDE, inside the card rather than the header (Jesse, 2026-09-21). Laid out as one wrapping
+   line of links rather than a list: six references stacked would make the card scroll past the
+   house rules it sits under, and these are a shelf to reach for, not a thing to read down. */
+#gamecardbody .guide{margin-top:10px;border-top:1px solid var(--line);padding-top:6px}
+#gamecardbody .guide p{margin:4px 0 0;font-size:12px;line-height:1.7;color:#8b94a3}
+#gamecardbody .guide a{color:#9fb6d8;text-decoration:none;border-bottom:1px solid #33404f}
+#gamecardbody .guide a:hover{color:#cfe0f5;border-bottom-color:#5aa9e6}
 /* An action you cannot take yet keeps its place but drops its light — the amber means "press me",
   so a disabled button must not wear it. */
 #actions button.blocked,#actions button:disabled{background:#232830;border:1px dashed #4a5361;
@@ -336,6 +343,27 @@ ul.blocked li{padding:2px 0}
 .lb-door:hover{border-color:#4d6fa8}
 .lb-door.active{background:#2f3a4b;border-color:#6f8fc8;color:#cfe0f5;font-weight:600}
 .lb-saved{display:flex;align-items:center;gap:10px}
+/* THE LOBBY'S ACTION BUTTONS CARRY THE SAME AMBER AS THE BOARD'S (Jesse, 2026-09-22).
+   `#actions button` has meant "this is a move you can make" all game; the lobby is the one screen
+   where a player is also being asked to act, and it was styling those buttons as plain chrome. The
+   selector is a LIST OF THE ACTIONS, not `#lobby button` — the settings form beneath Create is a
+   field of inputs, and amber on all of it would say everything is a move and so say nothing.
+   Disabled keeps `button:disabled` above: an unfilled Start game must not advertise a click. */
+#lb-look, #lb-join, #lb-create, #lb-start, #lb-copy, #lb-copylink, .lb-known-row .lb-rejoin{
+  background:#2b3444;border:2px solid #c8912f;box-shadow:0 0 0 1px rgba(200,145,47,.18);
+  color:#f2e6cf;font-weight:600}
+#lb-look:hover, #lb-join:hover, #lb-create:hover, #lb-start:hover, #lb-copy:hover,
+#lb-copylink:hover, .lb-known-row .lb-rejoin:hover{
+  background:#3a4a63;border-color:#f0b64a;box-shadow:0 0 0 3px rgba(240,182,74,.20)}
+#lb-look:focus-visible, #lb-join:focus-visible, #lb-create:focus-visible, #lb-start:focus-visible,
+#lb-copy:focus-visible, #lb-copylink:focus-visible, .lb-known-row .lb-rejoin:focus-visible{
+  outline:2px solid #f0b64a;outline-offset:2px}
+/* Copy is still secondary — it carries the amber so it reads as actionable, but not the weight of
+   the button that starts a game. `button.ghost` set its own border, so it is overridden here. */
+#lb-copy, #lb-copylink{border-width:1px;font-size:11px;box-shadow:none}
+/* A disabled action is chrome again, or the amber becomes a promise the button will not keep. */
+#lb-start:disabled, #lb-start:disabled:hover{
+  background:#2a3038;border:2px solid var(--line);box-shadow:none;color:inherit}
 .lb-known-row{display:flex;align-items:center;gap:10px;padding:6px 0;border-bottom:1px solid var(--line)}
 .lb-known-row:last-child{border-bottom:none}
 .lb-known-row .code{font-weight:700;letter-spacing:.06em;color:#f2e6cf;flex:1}
@@ -798,6 +798,44 @@ describe('Freight Agent operations (§6.3)', () => {
    );
  });

+  it('returns a cleared car to the yard EMPTY, not still loaded', () => {
+    /**
+     * REGRESSION, v0.8.1.0. `inboundCleared` pushed `pooled(e.stock)` under a comment reading
+     * "a car back in a yard is back in the common supply, carrying nothing" — and `pooled` does not
+     * do that. It strips the `origin` stamp and keeps `loaded` DELIBERATELY, because a train can
+     * retire at a Division Point with freight still aboard (`advance.ts`). So every car the Freight
+     * Agent cleared arrived in the Classification Yard still carrying a load that had already been
+     * delivered and already paid for.
+     *
+     * It bites hardest on coaches: `passengersDetrained` takes `type === 'coach' && !loaded` out of
+     * the Division Yard, and §2.2 refills the Division Yard from Classification — so a cleared
+     * coach came back as stock that could not unload a passenger.
+     */
+    const s = game();
+    const coord = withFacility(s);
+    const f = areaOf(s, 0).grid.get(coordKey(coord))!.facility!;
+    f.allows = { outbound: false, inbound: true };
+    f.inboundBox = [{ type: 'coach', loaded: true }];
+
+    const before = s.yards.classificationYard.length;
+    assert.ok(applyIntent(s, 0, { type: 'localOps.choose', option: 'freightAgent' }).ok);
+    assert.ok(
+      applyIntent(s, 0, { type: 'freightAgent.clearInbound', at: coord, index: 0 }).ok,
+      'the Freight Agent could not clear the red box',
+    );
+
+    assert.equal(f.inboundBox.length, 0, 'the red box was not cleared');
+    assert.equal(s.yards.classificationYard.length, before + 1, 'the car did not reach the yard');
+    const arrived = s.yards.classificationYard[s.yards.classificationYard.length - 1]!;
+    assert.equal(arrived.type, 'coach', 'the wrong car reached the yard');
+    assert.equal(arrived.loaded, false, 'the cleared car arrived in the yard still loaded');
+    // And a coach that arrives empty is stock a Porter can actually use again.
+    assert.ok(
+      s.yards.classificationYard.some((c) => c.type === 'coach' && !c.loaded),
+      'no empty coach reached the common supply',
+    );
+  });
+
  it('makes the Freight Agent option unavailable with nothing to operate (§6)', () => {
    // A player with no Facility cannot choose an option that has no possible follow-up.
    const s = game();
@@ -17,6 +17,7 @@ import { join } from 'node:path';

 import { fromSave, newGame, submit, toSave, view } from '../src/web/game.ts';
 import { actionGroups, currentActor } from '../src/web/game.ts';
+import { narrate } from '../src/sim/narrate.ts';

 const src = (rel: string): string => readFileSync(join(import.meta.dirname, '..', 'src', rel), 'utf8');

@@ -47,6 +48,12 @@ const KNOWN_UNREDUCED = [
  'clearanceRequested',
  'dispatchBonusUsed',
  'expediteFault',
+  /**
+   * The Freight Agent chose to do nothing (§6.3 requires no action). Emitted by `freightAgent.end`
+   * beside the `phaseEnded` that ends the turn, and reduces to nothing itself — exactly the
+   * `switchingEnded` pattern below.
+   */
+  'freightAgentIdled',
  /**
   * The New Train Phase's report that it could give a train nothing (playtest, 2026-09-16, the
   * Sparrow running empty). Emitted by the phase driver after the make-up round has nothing left to
@@ -161,3 +168,118 @@ describe('the intents are what reconstructs a game', () => {
    assert.deepEqual(Object.keys(toSave(game)).sort(), ['history', 'rules', 'seed']);
  });
 });
+
+/**
+ * A FREIGHT AGENT TURN, AS THE TABLE READS IT.
+ *
+ * Reported on Day 1 Stage 3 of v0.8.0.16, watching a bot: "one car moved to or from a facility —
+ * can we tell what car, what facility and whether it was to or from." Two separate faults sat
+ * behind that. The choice line ASSERTED a car had moved, which §6.3 does not require and the bot
+ * deliberately declines; and the three lines that do report the work named the grid coordinate
+ * rather than the industry standing on it.
+ */
+describe('a Freight Agent turn says what it did, to what, and where', () => {
+  const ctx = {
+    playerName: () => 'Bot 1',
+    facilityAt: (_p: number, c: { row: number; col: number }) =>
+      c.row === -1 && c.col === 1 ? 'the Freight House' : null,
+  };
+
+  it('announces the OPTION without claiming a car moved', () => {
+    const line = narrate({ type: 'localOpsOptionChosen', player: 0, option: 'freightAgent' } as never, ctx);
+    assert.ok(
+      !/one car moved/.test(line.text),
+      `the choice line still asserts an outcome: ${line.text}`,
+    );
+    // It must still say what the Freight Agent is FOR, or the option is a bare name.
+    assert.match(line.text, /Outbound|Inbound|jam/, `the choice line says nothing about the work: ${line.text}`);
+  });
+
+  it('says so when the Freight Agent deliberately does nothing', () => {
+    // The bot takes this route on purpose: unjamming a healthy box destroys a stocked load. Silence
+    // here is what made the turn read as a dropped turn.
+    const line = narrate({ type: 'freightAgentIdled', player: 0 } as never, ctx);
+    assert.match(line.text, /nothing worth doing/, `an idle Freight Agent is silent: ${line.text}`);
+  });
+
+  it('names the industry and the direction, not a coordinate', () => {
+    const at = { row: -1, col: 1 };
+    const stocked = narrate(
+      { type: 'stockToOutbound', player: 0, at, stock: { type: 'boxcar', loaded: true } } as never,
+      ctx,
+    );
+    assert.match(stocked.text, /the Freight House/, `no industry named: ${stocked.text}`);
+    assert.ok(!/\(1, -1\)/.test(stocked.text), `still speaking in coordinates: ${stocked.text}`);
+    assert.match(stocked.text, /boxcar/, `the car is not named: ${stocked.text}`);
+    assert.match(stocked.text, /INTO/, `the direction is not stated: ${stocked.text}`);
+
+    const cleared = narrate(
+      { type: 'inboundCleared', player: 0, at, stock: { type: 'hopper', loaded: true } } as never,
+      ctx,
+    );
+    assert.match(cleared.text, /the Freight House/, `no industry named: ${cleared.text}`);
+    assert.match(cleared.text, /OUT of/, `the direction is not stated: ${cleared.text}`);
+
+    const jam = narrate(
+      { type: 'facilityUnjammed', player: 0, at, from: 'menAtWork', stock: { type: 'tank', loaded: true } } as never,
+      ctx,
+    );
+    assert.match(jam.text, /the Freight House/, `no industry named: ${jam.text}`);
+
+    /**
+     * Where there is no industry, the coordinate is still the honest fallback rather than "nowhere".
+     *
+     * AND IT IS SPELLED THE WAY THE ACTION MENU SPELLS IT — X,Y, east/west then north/south. The
+     * log printed the internal row/col order until 2026-09-21, so the same square read "(1,-1)" in
+     * the menu and "(-1,1)" in the log, side by side. `test/web.test.ts` pins the two against each
+     * other; this pins the order itself.
+     */
+    const plain = narrate(
+      { type: 'stockToOutbound', player: 0, at: { row: -2, col: 4 }, stock: { type: 'boxcar', loaded: true } } as never,
+      ctx,
+    );
+    assert.match(plain.text, /\(4,-2\)/, `the fallback coordinate is not in X,Y order: ${plain.text}`);
+  });
+});
+
+/**
+ * WHICH MAINLINE CARD, AND WHAT THE CARD IS CALLED.
+ *
+ * Reported 2026-09-22, playing ABS Signals: "In the history, it referred to it as Mainline card 7,
+ * but didn't give the actual card type, which was plains. It should specify both. Note there were
+ * two plains cards dealt in this hand." Both halves matter for that reason — the name informs, the
+ * slot is the only thing telling two Plains apart.
+ */
+describe('the log names the Mainline card an Enhancement was built on', () => {
+  const ctx = {
+    mainlineAt: (node: number) => (node === 7 ? 'Plains' : node === 3 ? 'Heavy Grade' : null),
+    enhancementName: (key: string) => (key === 'absSignals' ? 'ABS Signals' : null),
+  };
+
+  it('gives the slot AND the card type', () => {
+    const line = narrate({ type: 'enhancementPlaced', player: 0, key: 'absSignals', node: 7 } as never, ctx);
+    assert.match(line.text, /Mainline card 7/, `the slot is gone: ${line.text}`);
+    assert.match(line.text, /Plains/, `the card type is missing: ${line.text}`);
+  });
+
+  it('calls the card by its printed name, not a split key', () => {
+    const line = narrate({ type: 'enhancementPlaced', player: 0, key: 'absSignals', node: 7 } as never, ctx);
+    assert.match(line.text, /ABS Signals/, `not the printed name: ${line.text}`);
+    assert.ok(!/abs Signals/.test(line.text), `still splitting the key: ${line.text}`);
+  });
+
+  it('still says something useful when the card cannot be resolved', () => {
+    // No resolver at all — an engine test narrating events has no division to ask.
+    const bare = narrate({ type: 'enhancementPlaced', player: 0, key: 'absSignals', node: 2 } as never, {});
+    assert.match(bare.text, /Mainline card 2/, `the slot must survive with no resolver: ${bare.text}`);
+  });
+
+  it('leaves a grid-square Enhancement naming its coordinate', () => {
+    const line = narrate(
+      { type: 'enhancementPlaced', player: 0, key: 'smallYard', at: { row: -1, col: 2 } } as never,
+      { enhancementName: (k: string) => (k === 'smallYard' ? 'Small Yard' : null) },
+    );
+    assert.match(line.text, /Small Yard/, `not the printed name: ${line.text}`);
+    assert.match(line.text, /\(2,-1\)/, `the square is gone or in the wrong order: ${line.text}`);
+  });
+});
@@ -59,6 +59,10 @@ const SAMPLES: GameEvent[] = [
  { type: 'deckReshuffled', order: ['c1', 'c2', 'c3', 'c4'], rngState: 7 },
  { type: 'departmentRefilled', slot: 0, cardId: 'c2' },
  { type: 'stockToOutbound', player: 0, at: { row: 1, col: 0 }, stock: { type: 'hopper', loaded: true } },
+  // Sampled from the start rather than joining the unsampled 25: an idle Freight Agent is the ONE
+  // line whose whole job is to be read when nothing happened, so an empty or fallback sentence
+  // would reproduce the silence it exists to fix.
+  { type: 'freightAgentIdled', player: 0 },
  { type: 'inboundCleared', player: 0, at: { row: 1, col: 0 }, stock: { type: 'hopper', loaded: true } },
  { type: 'facilityUnjammed', player: 0, at: { row: 1, col: 0 }, from: 'menAtWork', stock: { type: 'hopper', loaded: true } },
  { type: 'trainScheduled', player: 0, trainNumber: 4, roll: 7, slot: 6, rngState: 1 },
@@ -123,7 +127,7 @@ describe('narration', () => {
    /**
     * THE KNOWN GAP, PINNED SO IT CANNOT GROW.
     *
-     * `SAMPLES` exercises the TEXT of 30 of the 55 declared events; the other 25 have a narration
+     * `SAMPLES` exercises the TEXT of 31 of the 56 declared events; the other 25 have a narration
     * case (checked above) but no sample, so nothing proves their sentence is any good. Found
     * 2026-09-09 — the old test built both of its sets from `SAMPLES` and compared them to each
     * other, so it could only ever assert that the sample list had 30 distinct entries, and the one
@@ -230,7 +230,7 @@ describe('card catalogue (component 1)', () => {
     * "Both" column loses its only argument.
     *
     * Reported from playtesting v0.4.9d and confirmed by Jesse: the Refinery only ships tanks out,
-     * the Grocer's Warehouse only receives. `StationMaster-Home-Deck-v0.4.5.md` prints both that way,
+     * the Grocer's Warehouse only receives. `home-deck.md` prints both that way,
     * and so does the modifier set — all three Refinery modifiers grant outbound.
     */
    const houses = FREIGHT_PROFILES.filter(isFreightHouse).map((f) => f.kind);
@@ -562,8 +562,14 @@ describe('the history panel keeps the switching that matters', () => {
    assert.ok(submit(game, { type: 'switch.move', trayId, to: plain, reverse: false }), 'the move was refused');
    const moved = game.log.filter((l) => / moved the local crew/.test(l.text));
    assert.equal(moved.length, 1, `expected one move line, got ${moved.length}`);
-    // The FIRST move of a turn is kept, and this crew's first move is this one — so what is being
-    // checked here is that it names the square by what stands on it rather than by its coordinates.
-    assert.match(moved[0]!.text, /→ the Freight House|→ \(1,2\)/, `unexpected move line: ${moved[0]!.text}`);
+    /**
+     * The FIRST move of a turn is kept, and this crew's first move is this one — so what is being
+     * checked here is that it names the square by what stands on it rather than by its coordinates.
+     *
+     * The coordinate branch is `(2,1)` for `{ row: 1, col: 2 }`: X,Y, east/west then north/south,
+     * which is how `view.ts` has always written a square and how `narrate.ts` writes one since
+     * 2026-09-21. It read `(1,2)` here while the log still printed the internal storage order.
+     */
+    assert.match(moved[0]!.text, /→ the Freight House|→ \(2,1\)/, `unexpected move line: ${moved[0]!.text}`);
  });
 });
@@ -22,7 +22,7 @@ import { cardDescription, cardName, describeIntent, variantLabel } from '../src/
 import { variantsFor } from '../src/engine/track.ts';
 import { divisionSvg, officeSvg } from '../src/sim/board-svg.ts';
 import type { DivisionView } from '../src/sim/view.ts';
-import { ENHANCEMENT_RULES, STAGES_PER_DAY } from '../src/engine/content.ts';
+import { ENHANCEMENT_RULES, STAGES_PER_DAY, mainlineProfile } from '../src/engine/content.ts';
 import { dayEndHtml, facilitiesHtml, pilesHtml, resultsHtml, timetableHtml } from '../src/web/panels.ts';
 import { turnChartHtml } from '../src/sim/turnchart.ts';
 import { fieldSelectors } from '../src/web/settings-form.ts';
@@ -310,6 +310,119 @@ describe('the action menu presents choices the way they are made', () => {
    assert.deepEqual(shapes, ['en|ew', 'we|ws'], `the two rotations do not carry distinct shapes: ${shapes.join(' / ')}`);
  });

+  it('offers ABS Signals on EVERY Mainline card, each one named', () => {
+    /**
+     * REGRESSION — reported from a table on Day 1 Stage 1 of v0.8.0.16, and the THIRD instance of
+     * one trap. The action list drops duplicate labels, and `describeIntent` for a card play named
+     * the grid placement but never `node` — so every Mainline card produced the identical label
+     * "play ABS Signals" and all but the lowest-index one were discarded before the menu saw them.
+     * The card's own tooltip says "any Mainline card" while exactly one was ever on offer.
+     *
+     * The engine was never wrong: `check` accepts any node whose kind is 'mainline', and
+     * `legalActions` filters by `check`. The whole failure was in the label.
+     */
+    const game = newGame(555);
+    submit(game, actionGroups(game).options.find((o) => o.type === 'localOps.choose' && o.option === 'draw')!);
+
+    let absId: string | undefined;
+    for (const [id, card] of game.state.cards) {
+      const k = card.kind as { kind: string; key?: string };
+      if (k.kind === 'enhancement' && k.key === 'absSignals') { absId = id; break; }
+    }
+    assert.ok(absId, 'the deck has no ABS Signals card');
+    game.state.decks.hands.set(0, [absId]);
+
+    const mainlineNodes = game.state.division.nodes
+      .map((n, i) => ({ n, i }))
+      .filter(({ n }) => n.kind === 'mainline');
+    assert.ok(mainlineNodes.length > 1, 'this division has only one Mainline card — nothing to distinguish');
+
+    // The engine offers one per Mainline card ...
+    const offered = actionGroups(game).options.filter(
+      (o) => o.type === 'card.play' && o.cardId === absId && o.node !== undefined,
+    );
+    assert.equal(
+      offered.length,
+      mainlineNodes.length,
+      `the engine offers ${offered.length} placements for ${mainlineNodes.length} Mainline cards`,
+    );
+
+    // ... and every one of them must survive into the menu, which means distinct labels.
+    const labels = offered.map((o) => describeIntent(game.state, o));
+    assert.equal(
+      new Set(labels).size,
+      offered.length,
+      `the labels collapse, so the menu drops all but one: ${[...new Set(labels)].join(' / ')}`,
+    );
+
+    const spots = actionMenu(game)
+      .placeable.flatMap((g) => g.items)
+      .filter((it) => it.subjectKey === `card:${absId}`)
+      .flatMap((it) => it.spots);
+    assert.equal(
+      spots.length,
+      mainlineNodes.length,
+      `only ${spots.length} of ${mainlineNodes.length} Mainline cards can be chosen`,
+    );
+
+    /**
+     * And the card is called what the card face calls it. `prettyKey` rendered `absSignals` as
+     * "Abs Signals" on a button while the tooltip beside it said ABS — an acronym no key-splitter
+     * can recover, so the authored name in `ENHANCEMENT_CARDS` has to win.
+     */
+    assert.ok(
+      labels.every((l) => l.includes('ABS Signals')),
+      `the card is not called by its printed name: ${labels[0]}`,
+    );
+
+    // Each spot names the card it would go on, so the choice is legible rather than positional.
+    for (const { n } of mainlineNodes) {
+      const name = mainlineProfile((n as { card: Parameters<typeof mainlineProfile>[0] }).card).name;
+      assert.ok(
+        spots.some((sp) => sp.label.includes(name)),
+        `no spot names the ${name}: ${spots.map((sp) => sp.label).join(' / ')}`,
+      );
+    }
+  });
+
+  it('spells a square the same way in the action menu and in the log', () => {
+    /**
+     * REPORTED 2026-09-21. `view.ts` wrote "(col,row)" — X,Y, east/west then north/south, with a
+     * comment saying so — and `narrate.ts` wrote "(row,col)", the internal storage order, with no
+     * comment at all. So the action menu offered a move to "(1,-1)" and the log then reported it at
+     * "(-1,1)", in two panels a player reads side by side.
+     *
+     * PINNED AGAINST EACH OTHER rather than against a literal: a test asserting one format would
+     * have passed all along on whichever file it was written against. This compares the two
+     * renderers on the same square, which is the property that was actually broken.
+     */
+    const game = newGame(555);
+    const square = { row: -1, col: 2 };
+
+    const logged = describeIntent(game.state, {
+      type: 'switch.move',
+      trayId: [...game.state.trays.keys()][0]!,
+      to: square,
+      reverse: false,
+    });
+    const menu = describeIntent(game.state, {
+      type: 'card.play',
+      cardId: game.state.decks.hands.get(0)![0]!,
+      placement: square,
+    });
+
+    // Both must render the square, and render it identically.
+    const coord = /\((-?\d+,-?\d+)\)/;
+    const inLog = coord.exec(logged)?.[1];
+    const inMenu = coord.exec(menu)?.[1];
+    assert.ok(inLog, `the log line names no square: ${logged}`);
+    assert.ok(inMenu, `the menu line names no square: ${menu}`);
+    assert.equal(inMenu, inLog, 'the action menu and the log spell the same square differently');
+
+    // And the shared spelling is X,Y — east/west first, which is the order the map is drawn in.
+    assert.equal(inMenu, '2,-1', `not X,Y order: ${inMenu}`);
+  });
+
  it('says how deep a Department pile is, so a discard can be aimed', () => {
    // A discard goes ON TOP, so choosing where to put it is choosing whether to offer a card or to
    // bury one a rival wants. Neither is decidable without seeing what is already stacked up.
@@ -3010,6 +3123,82 @@ describe('the Division map shows the whole route', () => {
    return divisionSvg(snapshot(s, [], null).division);
  };

+  it('draws a signal on a Mainline card carrying ABS Signals, not only a tooltip', () => {
+    /**
+     * REPORTED FROM A TABLE, Day 1 Stage 1 of v0.8.0.16: "when played on the trestle, there was no
+     * on-the-card indication. It's only when you look at the tooltip for trestle that you see that
+     * ABS exists." The same complaint the Heavy Grade wedge below answers, and it matters more
+     * here — ABS is what decides whether running a second train onto that card is safe.
+     */
+    const s = createEngineGame({
+      id: 'div-abs',
+      seed: 7,
+      config: {
+        mode: 'competitive', days: 5, minCombinedRevenue: 0, maxCollisionsPerDay: 0, maxCollisionsTotal: 0,
+        pvpCardsAllowed: false,
+        optionalRules: { reducedVisibility: false, employeeRotation: false, emergencyToolbox: false },
+      },
+      playerNames: ['A', 'B'],
+    });
+
+    const before = divisionSvg(snapshot(s, [], null).division);
+    assert.ok(!before.includes('bs-abs'), 'a signal is drawn before ABS Signals was ever played');
+
+    const node = s.division.nodes.find((n) => n.kind === 'mainline');
+    assert.ok(node, 'this division has no Mainline card');
+    (node as { absSignals?: boolean }).absSignals = true;
+
+    const after = divisionSvg(snapshot(s, [], null).division);
+    assert.ok(after.includes('bs-abs-mast'), 'the card carrying ABS Signals draws no signal mast');
+    assert.ok(after.includes('bs-abs-lit'), 'the signal has no lit aspect');
+    // Exactly one card carries it, so the mark cannot be a row-wide decoration.
+    assert.equal((after.match(/bs-abs-mast/g) ?? []).length, 1, 'the signal is drawn on more than one card');
+    // And it stays in the tooltip too — the mark says THAT, the tip still says what it does.
+    assert.ok(after.includes('ABS Signals'), 'the tooltip stopped naming ABS Signals');
+  });
+
+  it('draws the Mainline modifiers on the card, not only in the tooltip', () => {
+    /**
+     * The other half of the ABS report (Jesse, 2026-09-21): "Brakeman, Airbrakes, Helpers and
+     * Realignment should also be drawn on the card, not just the tooltip."
+     *
+     * REALIGNMENT IS NOT IN THIS LIST ON PURPOSE. It never sits on a card — `reduce` takes the
+     * `became` branch and changes `node.card` outright — so a realigned card already announces
+     * itself by being a different card. Asserted below so the absence is a recorded finding rather
+     * than something that looks forgotten.
+     */
+    const s = createEngineGame({
+      id: 'div-mods',
+      seed: 7,
+      config: {
+        mode: 'competitive', days: 5, minCombinedRevenue: 0, maxCollisionsPerDay: 0, maxCollisionsTotal: 0,
+        pvpCardsAllowed: false,
+        optionalRules: { reducedVisibility: false, employeeRotation: false, emergencyToolbox: false },
+      },
+      playerNames: ['A', 'B', 'C', 'D'],
+    });
+
+    const node = s.division.nodes.find((n) => n.kind === 'mainline');
+    assert.ok(node, 'this division has no Mainline card');
+
+    assert.ok(!divisionSvg(snapshot(s, [], null).division).includes('bs-mod'), 'a tag is drawn with no modifier on');
+
+    (node as { modifiers?: string[] }).modifiers = ['brakeman', 'airbrakes', 'helpers'];
+    const svg = divisionSvg(snapshot(s, [], null).division);
+    for (const tag of ['BRK', 'AIR', 'HLP']) {
+      assert.ok(svg.includes(tag), `the ${tag} modifier is not drawn on the card`);
+    }
+    // And the tooltip still names them in full — the tag says THAT, the tip says WHICH.
+    assert.match(svg, /Brakeman/, 'the tooltip stopped naming the modifiers');
+
+    // Realignment converts the card instead of sitting on it, so it must never produce a tag.
+    (node as { modifiers?: string[] }).modifiers = ['realignment'];
+    assert.ok(
+      !divisionSvg(snapshot(s, [], null).division).includes('bs-mod'),
+      'Realignment drew a tag, but it changes the card rather than standing on it',
+    );
+  });
+
  it('draws which way a Heavy Grade climbs, instead of only saying it in the tooltip', () => {
    /**
     * REPORTED BY JESSE 2026-08-30: "heavy grade mainline card tooltip states climbs east, but card
@@ -5509,3 +5698,128 @@ describe('a Division Yard with no coaches is a reportable condition', () => {
    assert.equal(waiting.loaded + waiting.empty, 2, 'the waiting coaches were miscounted');
  });
 });
+
+// ---------------------------------------------------------------------------
+
+/**
+ * The Quickstart is published beside the game, so a tester on the box can reach it.
+ *
+ * WHY THIS IS A TEST. The link on the splash page is a plain href to a file the BUILD copies out of
+ * `docs/`. Nothing else connects the two: rename the document, or move it, and the build quietly
+ * publishes nothing while the splash page keeps offering a link that 404s. Neither `tsc` nor any
+ * other test would notice — the whole failure lives between a file name and a string.
+ */
+describe('the Quickstart guide reaches the site', () => {
+  it('is published into dist and linked from the splash page', () => {
+    const guide = join(dist, 'quickstart.md');
+    assert.ok(existsSync(guide), 'the build did not publish quickstart.md');
+
+    const text = readFileSync(guide, 'utf8');
+    assert.match(text, /^# Station Master — Quickstart/, 'quickstart.md is not the guide');
+    assert.match(
+      text,
+      /Describes the game as built at v/,
+      'the guide does not say which build it describes',
+    );
+
+    const splash = readFileSync(join(dist, 'index.html'), 'utf8');
+    assert.match(splash, /href="\.\/quickstart\.md"/, 'the splash page does not link the guide');
+  });
+
+  it('brings the lobby doors back every time the lobby is shown', () => {
+    /**
+     * REPORTED 2026-09-22: "After leaving a game, I go back to the lobby. I see the games that I'm
+     * in, and I have the ability to rejoin or forget. But what if I want to join a different game
+     * or start a new one? How do I get back there?"
+     *
+     * `enterSeating` hides `#lb-choice-section`, and until v0.8.1.0 only the lobby's OWN leave
+     * paths put it back. Leaving a RUNNING game is a third route — `main.ts` calls `runLobby` again
+     * — so the section stayed hidden from the previous seating and the screen came back holding
+     * nothing but "Games you are in", with the two doors on the page at `display:none` and no
+     * control that could reveal them.
+     *
+     * Read out of the source, like the MIME table above: the fix is an ordering property of one
+     * function, and there is no DOM harness here that drives `runLobby` end to end. What this pins
+     * is that the reset exists inside `runLobby` and happens BEFORE the resume path re-enters
+     * seating — reversing those two would hand a resuming browser the wrong screen.
+     */
+    const src = readFileSync(join(root, 'src/web/lobby.ts'), 'utf8');
+    const body = src.slice(src.indexOf('export function runLobby'));
+    assert.ok(body.length > 0, 'runLobby is gone');
+
+    const reset = body.indexOf("$('lb-choice-section').hidden = false");
+    assert.ok(reset > 0, 'runLobby never reveals the choice section, so leaving a game is a dead end');
+
+    const resume = body.indexOf('if (resume) enterSeating');
+    assert.ok(resume > 0, 'the resume path is gone');
+    assert.ok(reset < resume, 'the reset runs after the resume, which would show a seated browser the doors');
+
+    // And the seating section is put away with it, or the two would be on screen together.
+    const seating = body.indexOf("$('lb-seating-section').hidden = true");
+    assert.ok(seating > 0 && seating < resume, 'the seating section is not reset before resume');
+  });
+
+  it('publishes everything the guide links, so "Where to read more" is not five dead links', () => {
+    /**
+     * v0.8.0.16 published the Quickstart alone. Its §8 links five further documents by relative
+     * path, and every one of them 404'd on the package — verified against the running container,
+     * 5 of 6 paths missing. Publishing a guide without what it points at is the same broken-link
+     * failure as the test above, one hop further out, so it is pinned the same way: the links are
+     * read OUT OF THE GUIDE rather than listed here, or this test goes stale exactly as the
+     * references it guards did.
+     */
+    const guide = readFileSync(join(dist, 'quickstart.md'), 'utf8');
+    const section = guide.slice(guide.indexOf('## 8. Where to read more'));
+    assert.ok(section.length > 0, 'the guide no longer has a "Where to read more" section');
+
+    // Markdown links, minus anchors and absolute URLs — what a reader can actually click.
+    const targets = [...section.matchAll(/\]\(([^)#][^)]*)\)/g)]
+      .map((m) => m[1]!.replace(/^`|`$/g, ''))
+      .filter((t) => !/^https?:/.test(t));
+    assert.ok(targets.length >= 4, `only ${targets.length} references parsed out of the guide`);
+
+    for (const t of targets) {
+      assert.ok(existsSync(join(dist, t)), `the guide links ${t}, which the build does not publish`);
+    }
+  });
+
+  it('reaches the documentation from inside a game, in solitaire and multiplayer alike', () => {
+    /**
+     * Asked from a table, 2026-09-21: "how can we link the documentation so it can be reached from
+     * the gameplay, whether someone is playing solitaire or multiplayer?" The links live in the
+     * This Game card (Jesse's call) — and the point is that they need NO mode awareness, because
+     * both modes are the same page on the same origin. So this asserts the links exist and resolve,
+     * which is the whole of the mechanism.
+     *
+     * Read out of the built bundle rather than the source: what matters is what the shipped page
+     * offers, and a link that resolves in `src/` and not in `dist/` is the exact failure the two
+     * tests above exist to catch.
+     */
+    const bundle = readFileSync(join(dist, 'web', 'main.js'), 'utf8');
+    const guide = bundle.slice(bundle.indexOf('GUIDE_DOCS'), bundle.indexOf('GUIDE_DOCS') + 4000);
+    assert.ok(bundle.includes('GUIDE_DOCS') || bundle.includes('quickstart.md'), 'the bundle has no guide links');
+
+    // Every document offered in-game must be a file the build published.
+    const hrefs = [...guide.matchAll(/["'`](\.\/[A-Za-z0-9./-]+\.md)["'`]/g)].map((m) => m[1]!);
+    assert.ok(hrefs.length >= 5, `only ${hrefs.length} in-game guide links found`);
+    for (const h of hrefs) {
+      assert.ok(existsSync(join(dist, h.replace(/^\.\//, ''))), `the game links ${h}, which is not published`);
+    }
+
+    // A reference opened mid-turn must not take the game with it.
+    assert.ok(guide.includes('_blank'), 'the guide links would navigate away from a game in progress');
+    assert.ok(guide.includes('noopener'), 'a new-tab link without rel=noopener hands out a window handle');
+  });
+
+  it('is served as text rather than handed over as a download', () => {
+    /**
+     * The server's MIME fallback is `application/octet-stream`, which a browser downloads instead of
+     * displaying — so the link would hand a tester a file to save rather than a page to read. The
+     * table is read straight out of the source: asserting on a copy of it would pass while the real
+     * one was wrong.
+     */
+    const http = readFileSync(join(root, 'src/server/http.ts'), 'utf8');
+    const table = http.slice(http.indexOf('const MIME'), http.indexOf('const HEARTBEAT_MS'));
+    assert.match(table, /'\.md':\s*'text\/plain/, 'a .md file would be served as a download');
+  });
+});