Motir

Filed by the motir run MOTIR-3933 that halted on it. The corrected shape is submitted as a plan; this is the telemetry.

What the plan got wrong

MOTIR-3933 was authored by a run that had just found the defect on motir.co/design and had verified one half of it in a browser — adding data-appearance-scope to a live element and watching --el-accent move from the active palette's value to the tile's own. That half was real, and the card's estimate is a fair price for it: one attribute, one test, 3 points.

The card's other half — the base-entry leak — was described, not measured. It was written from a caveat paragraph in the component's own header, and sized by proximity to the half that had been measured.

Measured afterwards, by the run that claimed the card:

• The half that WAS reproduced is still incomplete: data-appearance-scope alone gives 9/10 in light and 2/10 in dark, because the nine per-palette dark blocks are compound [data-palette='X'][data-theme='dark'] selectors a nested scope cannot match. The card's own criterion 2 was therefore not satisfiable by the fix the card names.
• The half that was NOT reproduced is a different problem: 69 descendant material rules cross the nesting boundary, so under glassmorphism only 1 of 11 scoped style tiles wears its own material. Its honest remedy is a decision (selector rewrite vs isolated render tree), not a token block.

Why this one is worth recording

The mis-sizing is not carelessness — it is the predictable shape of a card written by the run that found the bug. A defect you have just reproduced comes with its remedy attached and its cost known; a defect you noticed beside it comes with neither, and it inherits the first one's estimate because they arrived in the same session and feel like one thing. The card even NAMED them as "Defect 1" and "Defect 2" and still carried one number.

It also survived validate_work_item. The over-sizing advisory fires at 13+ points or >70 minutes and reads the numbers the card SAYS; a card that is under-sized by an order of magnitude reports advisories: [] on the sizing axis and looks well-formed. Nothing mechanised could have caught it — only measuring the second remedy before writing the estimate.

Still open

• A publish + re-pin card is owed and does not exist. motir.co consumes the PUBLISHED @motir/design-system, so the fix reaches the live page only after a republish and a motir-marketing pin bump — the shape MOTIR-3872 used. That criterion was cut from MOTIR-3933 as a post-merge ordering violation, correctly, and cutting it left no card behind.

Acceptance criteria

• Closed when the submitted re-plan is approved or declined, and — if approved — when the publish + re-pin card above has been authored, so the deferral cut from MOTIR-3933 is not left as prose in a done card's body.
• No code change. This is a record.