Absurd trolley

A comic moral-choice runner that escalates from a simple fork in the tracks into delightfully stupid variants.

This is a browser game spec on Spekkr, published by nova. It is listed with the other specs. Paste the line to an agent that can open a web address, and the agent fetches the files and builds from them.

Spec
nova/absurd-trolley
Version
0.2.2
Published by
nova
Updated
15 Nov 2026

Build this spec

Build the spec at https://spekkr.com/nova/absurd-trolley/start.md.

Paste this to your AI agent. It opens the address, fetches the spec's files, records your choices and builds from them. Nothing to install, no account.

Published by nova.

What does the result look like?

Absurd trolley: heroAbsurd trolley: dessertAbsurd trolley: loop

What can I customize?

Script standard is 12 scenes long. short is 5, for a quicker run.

Your choices go into the line you copy. Anything else is yours to edit, but it leaves the tested path.

What do I get?

  • spec.md Requirements and acceptance checks.

    Absurd trolley

    What this builds

    A short branching story told as a series of scenes. Each scene shows a situation, two or three labelled choices, and a still. The first scene is a simple fork in a set of tracks. Later scenes twist the premise (the lever is a dessert menu, the tracks are email folders, the trolley is late for a meeting). Every scene and every branch is written in a script file shipped with the project. The writing is original.

    Read spekkr.inputs.json first. title is the first heading. player_name is substituted wherever the script has {name}. script chooses the long or short set. narrator and recap change what is shown around the scenes.

    Out of scope

    • A serious ethics textbook, scoring, or a claim that a choice is correct.
    • User-authored scenes in the first version.
    • Network, accounts or a shared ending board.

    Requirements

    Each requirement has an id. Checks name the id they cover.

    R1. Scenes
    • A scene has an id, a heading, a body, an image key, and two or three choices. Each choice has a label and the id of the next scene, or end.
    • The script file lists every scene. Missing next-ids are refused when the script is loaded.
    • {name} in any text is replaced with player_name.

    Accept when: loading each script succeeds, and a missing next-id in a fixture script is refused by name.

    R2. Branching
    • A choice moves to the documented next scene. There is no randomness in the path.
    • The first scene of standard and of short is fork.
    • A path is a list of (scene id, choice id). design.md lists three paths that must be followed exactly.

    Accept when: each documented path ends at the documented scene, and every step matches the table.

    R3. Chrome
    • If narrator is on, after the first choice a one-line aside from the chosen choice's aside field is shown under the next scene.
    • If recap is on, reaching end shows a list of every choice label in order, then a start-again control.
    • The first scene shows title as its heading, not the script's heading, for scene fork only.

    Accept when: path A with both options on shows the asides in the table and a recap of the three labels; with both off, neither appears.

    R4. Screenshots
    • The four frames in design.md are taken from path A at the named scenes.

    Accept when: they match the goldens within the tolerance in design.md.

    R5. Files
    • One command loads both scripts, walks the three paths, and writes the screenshots. No network.

    Accept when: the command passes from a clean folder.

    Verification

    Each scene's choices branch to the documented next scene. Three scripted paths (and one short-script path) are walked in tests and must end where the table says. Screenshots are taken at four points on path A. There is no randomness, no network, and no scoring of answers as right or wrong.

    Definition of done

    • Every acceptance check above runs from one command, and all pass at the person's inputs.
    • A README says how to play, how to run the checks, and how to edit a scene.
    # Absurd trolley
    
    ## What this builds
    
    A short branching story told as a series of scenes. Each scene shows a
    situation, two or three labelled choices, and a still. The first scene is
    a simple fork in a set of tracks. Later scenes twist the premise (the
    lever is a dessert menu, the tracks are email folders, the trolley is late
    for a meeting). Every scene and every branch is written in a script file
    shipped with the project. The writing is original.
    
    Read `spekkr.inputs.json` first. `title` is the first heading.
    `player_name` is substituted wherever the script has `{name}`. `script`
    chooses the long or short set. `narrator` and `recap` change what is shown
    around the scenes.
    
    ## Out of scope
    
    - A serious ethics textbook, scoring, or a claim that a choice is correct.
    - User-authored scenes in the first version.
    - Network, accounts or a shared ending board.
    
    ## Requirements
    
    Each requirement has an id. Checks name the id they cover.
    
    ### R1. Scenes
    
    - A scene has an id, a heading, a body, an image key, and two or three
      choices. Each choice has a label and the id of the next scene, or `end`.
    - The script file lists every scene. Missing next-ids are refused when the
      script is loaded.
    - {name} in any text is replaced with `player_name`.
    
    Accept when: loading each script succeeds, and a missing next-id in a
    fixture script is refused by name.
    
    ### R2. Branching
    
    - A choice moves to the documented next scene. There is no randomness in
      the path.
    - The first scene of `standard` and of `short` is `fork`.
    - A path is a list of (scene id, choice id). `design.md` lists three paths
      that must be followed exactly.
    
    Accept when: each documented path ends at the documented scene, and every
    step matches the table.
    
    ### R3. Chrome
    
    - If `narrator` is on, after the first choice a one-line aside from the
      chosen choice's `aside` field is shown under the next scene.
    - If `recap` is on, reaching `end` shows a list of every choice label in
      order, then a start-again control.
    - The first scene shows `title` as its heading, not the script's heading,
      for scene `fork` only.
    
    Accept when: path A with both options on shows the asides in the table and
    a recap of the three labels; with both off, neither appears.
    
    ### R4. Screenshots
    
    - The four frames in `design.md` are taken from path A at the named
      scenes.
    
    Accept when: they match the goldens within the tolerance in `design.md`.
    
    ### R5. Files
    
    - One command loads both scripts, walks the three paths, and writes the
      screenshots. No network.
    
    Accept when: the command passes from a clean folder.
    
    ## Verification
    
    Each scene's choices branch to the documented next scene. Three scripted
    paths (and one short-script path) are walked in tests and must end where
    the table says. Screenshots are taken at four points on path A. There is
    no randomness, no network, and no scoring of answers as right or wrong.
    
    ## Definition of done
    
    - Every acceptance check above runs from one command, and all pass at the
      person's inputs.
    - A README says how to play, how to run the checks, and how to edit a
      scene.
    

    Open spec.md as plain text

  • design.md The scene scripts, the documented paths and the screenshot points.

    Design

    This file fixes the parts that are easy to get wrong. The language and tools are up to you, as long as the numbers here hold.

    Scripts

    scripts/standard.json has 12 scenes, scripts/short.json has 5. Both start at fork. Every choice points at a scene id in the same file or at end.

    Scene fork (both scripts), body trimmed here:

    • Heading used only if title is empty: "A fork".
    • Choices: left "Pull the lever" → dessert; right "Leave the lever" → email.

    The remaining scenes, their exact text, and their choices are in those files. Tests read the files. A change to a next-id must be made in the file and in the path table together.

    Documented paths (standard)

    PathSteps (scene, choice)Ends at
    A(fork, left), (dessert, cake), (meeting, wait)recap then end if recap is on, else end
    B(fork, right), (email, archive), (folders, both)end
    C(fork, left), (dessert, pie), (meeting, run)end

    The short script has path S: (fork, left), (dessert, cake) → end.

    Path A asides, when narrator is on:

    After choiceAside shown on the next scene
    left"A decisive tug."
    cake"Sweetness, at least."
    wait"The trolley, and the meeting, go on without you."

    Golden frames (path A, standard, both options on)

    FileWhen
    first-forkScene fork, nothing chosen yet
    chosen-leftScene dessert, after left
    escalationScene meeting
    recapThe recap screen

    Tolerance: each channel within 8 levels, 98% of pixels.

    Inputs

    KeyLabelDefaultNotes
    titleTitleA small problem of tracksShown on the first scene.
    player_nameName on the leverAlexUsed in later scenes when the text addresses the person.
    narratorNarrator asidesonA second line under the scene that comments on the last choice.
    recapEnd recaponA last screen lists every choice in the path.
    scriptScriptstandardstandard is 12 scenes long. short is 5, for a quicker run.

    Acceptance at the defaults

    The numbers, names and fixture paths already in this file are the authority. Where a later table or a shipped fixture file disagrees with prose, the table or the fixture wins.

    FixtureRole
    design.mdA shipped fixture. The build may move it, but then design.md must name the new path.
    scripts/short.jsonA shipped fixture. The build may move it, but then design.md must name the new path.
    scripts/standard.jsonA shipped fixture. The build may move it, but then design.md must name the new path.

    Golden frames are compared at 1400 by 1000 CSS pixels unless a requirement names another size. Tolerance, unless a tighter figure is given above: 8 levels, 98% of pixels, with motion, cursors and clocks frozen as this file says. A build's goldens are of that build; they do not have to match a shipped reference preview pixel-for-pixel if the tool or the type is different, but they do have to match this file's layout, counts and copy.

    Screenshots

    These pictures are headless-Chrome captures of the working reference in reference/preview.html, at the default inputs. They are not drawings of a product that has not been built. The first picture is the one Spekkr shows on the spec's card. A build may ship its own goldens; those goldens are of that build and do not have to match these frames pixel-for-pixel.

    FileWhat it shows
    shots/tracks.pngThe trolley approaches a fork: uneaten bread one way, a violin the other.

    How to verify

    1. Run the checks at the default inputs, without a network.
    2. Walk through the scripted playthroughs, fixture lookups or goldens named in the requirements; each must match this file.
    3. Repeat once at the person's inputs when they differ from the defaults.
    4. Each scene's choices branch to the documented next scene. Three scripted paths (and one short-script path) are walked in tests and must end where the table says. Screenshots are taken at four points on path A. There is no randomness, no network, and no scoring of answers as right or wrong.
    # Design
    
    This file fixes the parts that are easy to get wrong. The language and
    tools are up to you, as long as the numbers here hold.
    
    ## Scripts
    
    `scripts/standard.json` has 12 scenes, `scripts/short.json` has 5. Both
    start at `fork`. Every choice points at a scene id in the same file or at
    `end`.
    
    Scene `fork` (both scripts), body trimmed here:
    
    - Heading used only if `title` is empty: "A fork".
    - Choices: `left` "Pull the lever" → `dessert`; `right` "Leave the lever"
      → `email`.
    
    The remaining scenes, their exact text, and their choices are in those
    files. Tests read the files. A change to a next-id must be made in the
    file and in the path table together.
    
    ## Documented paths (standard)
    
    | Path | Steps (scene, choice) | Ends at |
    | --- | --- | --- |
    | A | (fork, left), (dessert, cake), (meeting, wait) | `recap` then `end` if `recap` is on, else `end` |
    | B | (fork, right), (email, archive), (folders, both) | `end` |
    | C | (fork, left), (dessert, pie), (meeting, run) | `end` |
    
    The short script has path S: (fork, left), (dessert, cake) → `end`.
    
    Path A asides, when `narrator` is on:
    
    | After choice | Aside shown on the next scene |
    | --- | --- |
    | left | "A decisive tug." |
    | cake | "Sweetness, at least." |
    | wait | "The trolley, and the meeting, go on without you." |
    
    ## Golden frames (path A, standard, both options on)
    
    | File | When |
    | --- | --- |
    | `first-fork` | Scene `fork`, nothing chosen yet |
    | `chosen-left` | Scene `dessert`, after `left` |
    | `escalation` | Scene `meeting` |
    | `recap` | The recap screen |
    
    Tolerance: each channel within 8 levels, 98% of pixels.
    
    ## Inputs
    
    | Key | Label | Default | Notes |
    | --- | --- | --- | --- |
    | `title` | Title | A small problem of tracks | Shown on the first scene. |
    | `player_name` | Name on the lever | Alex | Used in later scenes when the text addresses the person. |
    | `narrator` | Narrator asides | on | A second line under the scene that comments on the last choice. |
    | `recap` | End recap | on | A last screen lists every choice in the path. |
    | `script` | Script | standard | standard is 12 scenes long. short is 5, for a quicker run. |
    
    ## Acceptance at the defaults
    
    The numbers, names and fixture paths already in this file are the
    authority. Where a later table or a shipped fixture file disagrees with
    prose, the table or the fixture wins.
    
    | Fixture | Role |
    | --- | --- |
    | `design.md` | A shipped fixture. The build may move it, but then `design.md` must name the new path. |
    | `scripts/short.json` | A shipped fixture. The build may move it, but then `design.md` must name the new path. |
    | `scripts/standard.json` | A shipped fixture. The build may move it, but then `design.md` must name the new path. |
    
    Golden frames are compared at 1400 by 1000 CSS pixels unless a requirement
    names another size. Tolerance, unless a tighter figure is given above: 8
    levels, 98% of pixels, with motion, cursors and clocks frozen as this file
    says. A build's goldens are of that build; they do not have to match a
    shipped reference preview pixel-for-pixel if the tool or the type is
    different, but they do have to match this file's layout, counts and copy.
    
    ## Screenshots
    
    These pictures are headless-Chrome captures of the working reference in
    `reference/preview.html`, at the default inputs. They are not drawings of a
    product that has not been built. The first picture is the one Spekkr shows on
    the spec's card. A build may ship its own goldens; those goldens are of that
    build and do not have to match these frames pixel-for-pixel.
    
    | File | What it shows |
    | --- | --- |
    | `shots/tracks.png` | The trolley approaches a fork: uneaten bread one way, a violin the other. |
    
    
    ## How to verify
    
    1. Run the checks at the default inputs, without a network.
    
    2. Walk through the scripted playthroughs, fixture lookups or goldens
    named in the requirements; each must match this file.
    
    3. Repeat once at the person's inputs when they differ from the defaults.
    
    4. Each scene's choices branch to the documented next scene. Three
    scripted paths (and one short-script path) are walked in tests and must
    end where the table says. Screenshots are taken at four points on path A.
    There is no randomness, no network, and no scoring of answers as right or
    wrong.
    

    Open design.md as plain text

  • tasks.md Build tasks in order, each with a check.

    Tasks

    Work in order. Finish a task's check before starting the next.

    1. Set up. Set up a web app (or the stack you have chosen) and a test runner that can open a page, click, and compare screenshots. Read spekkr.inputs.json first. Check: a blank or heading page loads, and an empty test run passes.
    2. Scenes (R1). Build what R1. Scenes asks for, to the numbers in design.md. In particular: A scene has an id, a heading, a body, an image key, and two or three choices. Each choice has a label and the id of the next scene, or end. Check: the R1. Scenes acceptance checks pass.
    3. Branching (R2). Build what R2. Branching asks for, to the numbers in design.md. In particular: A choice moves to the documented next scene. There is no randomness in the path. Check: the R2. Branching acceptance checks pass.
    4. Chrome (R3). Build what R3. Chrome asks for, to the numbers in design.md. In particular: If narrator is on, after the first choice a one-line aside from the chosen choice's aside field is shown under the next scene. Check: the R3. Chrome acceptance checks pass.
    5. Screenshots (R4). Build what R4. Screenshots asks for, to the numbers in design.md. In particular: The four frames in design.md are taken from path A at the named scenes. Check: the R4. Screenshots acceptance checks pass.
    6. Files (R5). Build what R5. Files asks for, to the numbers in design.md. In particular: One command loads both scripts, walks the three paths, and writes the screenshots. No network. Check: the R5. Files acceptance checks pass.
    7. Goldens. Capture the frames named in design.md at the defaults, with motion frozen where the file says so. Compare them within the stated tolerance. Check: the frames match, and a pixel-level check fails if a documented element is removed on purpose.
    8. Finish. Write a README that says how to install what is needed, how to change the inputs, and how to run the checks. Run the one command from a clean folder. Check: every item under Definition of done in spec.md is true.
    # Tasks
    
    Work in order. Finish a task's check before starting the next.
    
    1. **Set up.** Set up a web app (or the stack you have chosen) and a test
    runner that can open a page, click, and compare screenshots. Read
    `spekkr.inputs.json` first.
       Check: a blank or heading page loads, and an empty test run passes.
    
    2. **Scenes (R1).** Build what `R1. Scenes` asks for, to the numbers in
    `design.md`. In particular: A scene has an id, a heading, a body, an image
    key, and two or three choices. Each choice has a label and the id of the
    next scene, or `end`.
       Check: the `R1. Scenes` acceptance checks pass.
    
    3. **Branching (R2).** Build what `R2. Branching` asks for, to the numbers
    in `design.md`. In particular: A choice moves to the documented next
    scene. There is no randomness in the path.
       Check: the `R2. Branching` acceptance checks pass.
    
    4. **Chrome (R3).** Build what `R3. Chrome` asks for, to the numbers in
    `design.md`. In particular: If `narrator` is on, after the first choice a
    one-line aside from the chosen choice's `aside` field is shown under the
    next scene.
       Check: the `R3. Chrome` acceptance checks pass.
    
    5. **Screenshots (R4).** Build what `R4. Screenshots` asks for, to the
    numbers in `design.md`. In particular: The four frames in `design.md` are
    taken from path A at the named scenes.
       Check: the `R4. Screenshots` acceptance checks pass.
    
    6. **Files (R5).** Build what `R5. Files` asks for, to the numbers in
    `design.md`. In particular: One command loads both scripts, walks the
    three paths, and writes the screenshots. No network.
       Check: the `R5. Files` acceptance checks pass.
    
    7. **Goldens.** Capture the frames named in `design.md` at the defaults,
    with motion frozen where the file says so. Compare them within the stated
    tolerance.
       Check: the frames match, and a pixel-level check fails if a documented
    element is removed on purpose.
    
    8. **Finish.** Write a README that says how to install what is needed, how
    to change the inputs, and how to run the checks. Run the one command from
    a clean folder.
       Check: every item under Definition of done in `spec.md` is true.
    

    Open tasks.md as plain text

  • reference/preview.html A working reference preview of the default view, used to make the listing pictures. A build does not have to start from this file.
  • reference/fonts/fonts.css Webfont faces for the reference preview.
  • reference/fonts/Geist.ttf Geist variable typeface for the reference preview.
  • reference/fonts/Newsreader.ttf Newsreader variable typeface for the reference preview.
  • reference/fonts/Newsreader-Italic.ttf Newsreader italic variable typeface for the reference preview.
  • reference/fonts/InstrumentSerif.ttf Instrument Serif for the reference preview.
  • reference/fonts/InstrumentSerif-Italic.ttf Instrument Serif italic for the reference preview.

Is it for me?

Platforms
web
Tools
TypeScript

What does it need?

Agent permissions
shell
Commands it runs
install a test runner install a tool that compares screenshots

What changed?

  • 0.2.215 Nov 2026Names what it builds.
  • 0.2.115 Nov 2026Give the trolley real volume, better bread and violin props, and an animated loop.
  • 0.2.014 Nov 2026Polished reference preview and new listing shots, including a short loop.
  • 0.1.014 Nov 2026First version.

For agents: start.md says how to build this spec, spec.json holds everything on this page, and bundle.tar.gz holds its files.