Module activity and spend history

Every module’s own page keeps a running record of what’s happened to it and what it has cost — not just the version you have today. The Activity tab is where both live: a timeline of every version, import, and enhancement, and a spend breakdown that includes attempts that never shipped anything.

This page covers the timeline and the spend breakdown. Reports you’ve filed against the module also appear here, but sharing feedback already covers reading and filing those — see its “After you submit” section.

Opening the Activity tab

1. Open the Modules Library and click a module’s name to open its own page.
2. Click the Activity tab.

A module with no history yet — never enhanced, imported, or reported on — shows Nothing has happened on this module yet. instead of an empty timeline.

The timeline

The timeline lists, newest first, every event recorded against the module:

• v_N_ promoted — a version became current, whether it came from a generation, an enhancement, or a rollback. A rollback shows up the same way as any other promotion: as the version it produced, not as a distinct kind of event.
• v_N_ imported — an imported ZIP became a version.
• Attempt failed — a run that didn’t produce a version, with the error underneath.
• Nothing to change — an enhancement run that finished without changing anything.
• Enhancement started — an enhancement session began. There’s no separate “enhancement finished” event; a successful one shows up as the v_N_ promoted event for the version it produced instead.
• You reported a problem — a feedback report you filed, with its GitHub issue number and status if it has one.

Most events carry a small status chip. Versions and imports show Ready, Needs attention, or Blocked — the same states as the module’s own status pill, but for that specific version. Attempts show Failed or No change. An enhancement’s start shows a chip once its session reaches a resolved state — Promoted, Failed, No change, or Abandoned — but while it’s still being discussed, reviewed, or built, the event sits with no chip yet.

Note

Where an event has more detail — an enhancement’s request text, or a failed attempt’s error — it’s shown underneath, trimmed to a short excerpt.

Filtering the timeline

A row of chips above the timeline narrows it down: All, Versions, Enhancements, Reports. Versions also includes failed and no-change attempts, not just promotions and imports — there’s no separate filter for attempts on their own.

Spend on this module

Next to the timeline, the Spend on this module card shows the total cost across the module’s whole lifetime — every version, every enhancement session, every failed attempt, and the conversation that first refined the requirement into a spec. That’s a wider number than any one version’s own cost: it includes money spent on runs that never shipped anything.

If any spend is broken down by step, a table lists each one as Step · Cost · Time. The step name combines what the spend was for with which version it belongs to:

• v_N_ — spend tied to a specific shipped version: its generation cost, its browser tests, or — once an enhancement’s result was promoted — that enhancement’s own review, discussion, and build cost. An imported version gets an Import (analysis + install) row too, currently always $0.00 — importing doesn’t run the AI agent, and the row exists to show the step happened, not to charge for it.
• Current enhancement — spend on an enhancement session that’s still going. A failed attempt’s spend row is labelled this way too, even once the attempt is long over — for that row, go by the Failed attempt stage next to it and the matching Attempt failed event in the timeline, not by this label, to tell whether it’s finished.
• Not promoted — an enhancement session that ended (failed or abandoned) without shipping a version.
• Module — the refinement conversation that produced the spec in the first place. It isn’t tied to any one version, so it’s charged to the module as a whole.

The stage after that names the kind of work: Refinement, Review, Discussion, Build & verify, Generation, Import (analysis + install), Failed attempt, or Browser tests. An enhancement’s own cost splits across its review, discussion, and build stages rather than showing as one lump sum — that’s why an enhanced version’s row list can look longer than a first-generated version’s.

The Time column shows how long that step took (for example 45s, 3m 12s, or 1h 5m) when a wall-clock duration was recorded, running while it’s still going, or — when neither applies — which is always the case for the refinement row, since a conversation that produced a spec is necessarily over by the time there’s a module page to look at.

Note

An enhancement run that finished with nothing to change doesn’t add its own line to this breakdown, even though its Nothing to change event is in the timeline above. Only a failed unpromoted attempt gets a spend row of its own.

If nothing has been spent in a way that breaks down by step, the card shows just the total, with no table underneath.