Malatang prep spec-card SOP - one measurable spec per item, new hires cut to the card, spot-checks run with a ruler

You are the kitchen-standardization consultant for a malatang shop. I am the head chef,
building a spec-card SOP for the prep station so cut quality no longer depends on one
person being present. Produce the full SOP and card templates within this framework; I
fill in my shop's numbers wherever you see [ ].

SOP title: Prep Spec Card and Spot-Check Standard
Applies to: prep cooks (including temporary cover), chef on duty

1. The spec card (core asset - one card per item)
Card fields:
- Item name and photo (an actual photo, not a drawing)
- Cut form: segment / slice / cube / strip
- Measurable spec: segment length [cm], slice thickness [mm], cube edge [cm] -
  must be measurable; "medium" or "a bit thinner than potato" is invalid and gets
  sent back for rewrite
- Post-soak weight standard (dried items): [g] per portion after soaking
- Basket quantity: [jin / portions] per basket
- Draining standard: drain [minutes], or an observable action ("three shakes,
  no dripping")
- Destination: which fridge bin it stocks
Issuance rule: specs are set and signed by the head chef with a date; changes
need a signature too; prep cooks may not alter specs on their own

2. Onboarding (a new hire's first day)
- New hires don't learn specs by watching - they cut to the card: first three
  items against the card, the veteran corrects with the ruler, then the range
  expands
- Three-day test: five random items pulled from the cards; four or more in
  spec passes

3. Spot-checks (daily)
- After opening setup, the chef on duty pulls [3] random items and checks
  against the card with ruler / scale
- Pass: all three sampled items in spec
- Fail: that item's batch gets reworked (re-cut or downgraded to staff meal)
  and one deviation gets logged on the card
- The same item deviating [3] days running -> not a person problem, a card
  or material problem: check whether the spec needs updating or the incoming
  produce changed

4. Spec-review triggers (any one forces re-measuring the spec)
1. New supplier or origin (moisture and density differ - same spec, different
   doneness)
2. New knife or cutting equipment
3. Customer feedback of the "undercooked / mushy / can't pick it up" family
4. Season change (when the same item's moisture visibly shifts between
   spring and autumn)

Output format:
- SOP body in the four sections above, each with owner / timing / standard /
  failure path
- One blank "prep spec card" template (all fields) and one filled example card
  (use potato slices; numbers left as [ ])
- One "daily spot-check log" template: date / item pulled / measured result /
  ruling / action, one line per item

Constraints:
- All example numbers stay as [ ] for me to fill - no "universal cutting specs"
- The words "appropriate amount," "as needed," and "by experience" never appear
  on a card - any one of them voids it
- The card is not a training manual (no why-paragraphs on the card - fields and
  numbers only; explanations live in the SOP body)
- Spot-checks happen by ruler, not by eye - "looks about right" is not a check

Why it is written this way

Malatang is an order of magnitude more sensitive to cut spec than stir-fry cooking: everything blanches in the same broth on roughly the same clock, and that "roughly the same" only works if the specs are roughly the same - a doubled-thick potato slice keeps a raw core under identical blanching, and an unflattened cabbage rib sits crunchy next to overcooked leaves. A fridge of uneven specs hands doneness over to chance. Hence this SOP's first principle: specs must be measurable. "Medium" is ruled invalid because an unmeasurable standard is no standard - it can't be spot-checked and can't settle a dispute.

Separating the card from the teaching material (fields and numbers on the card, why in the body) is a rule earned in practice: the card hangs at the prep table, and the worker mid-task must be able to grab the number at a glance. Insert a paragraph on cutting theory and the cost of finding a number doubles - after which nobody consults the card. Same reason the example card insists on a real photo over a drawing: drawings flatter, photos don't.

Spot-checks run "by ruler, not by eye" because eyeball standards drift with the day: busy today, close-enough passes; idle tomorrow, the same cut gets sent back. A ruler doesn't drift. And "three consecutive deviations on one item means check the card or the material" matters most of all - it flips the spot-check from a people-catching tool into a sensor for system change: when incoming moisture shifts or a supplier swaps, the first alarm is repeated deviation on the same card. Log every deviation against a person and nobody ever looks at the system.

The seasonal trigger stands alone among the four because it's the most invisible: the same vegetable can differ materially in moisture between spring and autumn - the cut spec unchanged, the doneness changed. A veteran senses "this batch is off," but without a card and a re-measure rule, the sense never becomes an action.

Variables