Deriving verdicts from stored observations…
price · 85% confidence in this cause
Both the value and the element it was read from changed.
nothing extractedv1price stopped producing a value; it last read successfully at 2026-08-12 from div.product > div.pricing > span.price. No candidate in the plan matches the page any more.
Repair from the digest captured with the failing run. It holds ranked replacement candidates gathered at the moment of failure, so this costs no second page load.
Each run judged as it arrived. A marked row is one where the judges disagreed.
| Run | Completed | Value | Extraction | Cause | Plan |
|---|---|---|---|---|---|
| 2026-08-13 | failed2/3 | implausible | failed | element-removed | v1 |
| 2026-08-12 | ok3/3 | unverified | trusted | healthy | v1 |
| 2026-08-11 | ok3/3 | unverified | trusted | healthy |
No replacement selector is supported by the recorded evidence, so no repair is proposed. A confident cause with nothing to propose escalates — it does not guess.
| v1 |
| 2026-08-10 | ok3/3 | unverified | trusted | healthy | v1 |
The evidence the verdicts above were derived from. Signals implying everything is fine are not listed.
extraction-lostprice has produced no value for 1 consecutive run; it last read successfully at 2026-08-12 09:00:00 from div.product > div.pricing > span.price.
Read the structural digest captured with the failing run — it holds ranked replacement candidates gathered at the moment of failure, so diagnosis costs no second page load.
value-missingprice produced no value.
required-field-missingprice is required by plan v1 and produced no value.
Check the stored structural digest for this run — candidate replacements were captured at the moment it failed, so no second page load is needed.
value-lostprice produced no value, after 3 consecutive observations that did.
This is a break rather than an absence. The structural digest for this run holds the candidate replacements.