Braised-food scheduling supervisor - the pre-release roster self-check: skill coverage, hours red lines, absence backup, rotation fairness - tick every box before it hits the group chat

Here is a user need. Expand it into a "Roster Pre-Release Self-Check"
for a braised-food store's scheduling supervisor to run before the
roster goes to the staff group chat. Mostly tables, checkbox per item,
concrete fail examples.

Background (put it in the header):
- Store roles: braising (morning), chopping, register, stocking -
  different skill levels, part-time and full-time mixed
- Roster cycle: [weekly], published [3] days ahead
- The self-check runs before release; fix everything first - a
  published roster should not need changes

Five check groups (each item states "what to check" and "what fail
looks like"):

Group 1, Skill coverage (walk every daypart):
- Every trading daypart has at least 1 person who can chop alone and
  1 who can run the register alone on shift. Fail = a daypart staffed
  entirely with register-only skills, chopping waits for someone
  else's arrival
- The braising morning shift includes someone who can "mind the pot"
  (holds keys or knows master-brine upkeep). Fail = two trainees
  rostered in the morning, nobody tends the brine
- Every critical skill (pot / chopping / solo register) is held by
  at least 2 people - one holder = a single point. Fail = a skill
  only one person in the store has, and that person is rostered
  every day this week without a day off

Group 2, Hours red lines (walk every person, against the written
rules - not memory):
- Monthly total hours and overtime within local legal limits.
  Fail = anyone over the line, even by one hour
- Rest interval between consecutive shifts meets local rules.
  Fail = closing shift followed by next-day opening (insufficient
  gap)
- Weekly rest days meet local rules. Fail = someone going [n]
  consecutive weeks without a full day off
- Student part-timers' hours don't collide with their declared
  unavailability. Fail = rostered into a slot they filed off
  in advance

Group 3, Absence and backup:
- Approved leave dates genuinely carry no shifts. Fail = a shift on
  an approved leave day (the number-one complaint source)
- A sudden-absence backup exists: every peak daypart names one
  on-call cover. Fail = blank backup column

Group 4, Rotation fairness:
- Weekend rotation record: no one person carries all weekend shifts
  [2] weeks running. Fail = weekends repeatedly landing on the
  same person
- Holiday shifts have a rotation record or compensation plan.
  Fail = holidays always worked by the same agreeable few
- Closing and opening shifts don't permanently belong to one person.
  Fail = someone rostered close-only for months (the worst shift
  for commutes and life)

Group 5, Peak-season flex:
- Pre-holiday weeks scheduled separately from ordinary weeks.
  Fail = pre-holiday roster identical to an ordinary week
- Reserved on-call part-time hours: [x] per week, stated in the
  roster notes. Fail = zero flex - one absence and the store runs
  uncovered

Three fixed closing lines:
- Self-checker's signature / date
- One sentence on "what this release changed" (on record; next
  check verifies those spots first)
- The "no post-release edits" pledge: when a change is truly needed,
  explain it in the group chat and tag everyone affected - making
  the cost of editing visible

Expansion requirements:
- All [] values left blank for the store and local rules to fill
- Every hours red line points to "check against the written rule
  text" - the checklist itself prints no specific hour counts;
  they vary by place and period, and hard-coding them plants a mine
- No "communicate more" items - nothing that can't be ticked

Why it is written this way

A roster's credibility is a single-use asset: revise it after it hits the group chat, and from the second time on, staff read every roster as a draft and hold off confirming personal plans - doubling the scheduler's weekly workload. This checklist exists to move every edit before release, and the "no post-release edits" pledge is a shackle the supervisor buckles on themselves: making the cost of editing visible (explain in the group, tag the affected) is what actually buys those extra ten minutes of care before sending.

Group 2 deliberately prints no numbers and points every check at "the written rule text" - the checklist's most counterintuitive design. The scheduling supervisor's most common violation isn't willful; it's executing a remembered number that's outdated or misremembered. If the checklist dares to print "overtime under X hours," the supervisor will dare to skip opening the source. Print "check the text" and the source actually gets opened this week. The "declared unavailability" check sits in this group on purpose: in form it's a scheduling slip, in substance it's a trust collapse - one shift on an approved leave day is the number-one complaint source and can undo the seriousness of the whole leave-filing system in one stroke.

Group 1's three items map the three single points peculiar to a braised-food shop: a daypart with nobody who can chop (customers pick, nobody cuts), a braising morning with nobody who can mind the pot (the master brine sits unattended overnight - an asset-level risk), and a skill held by exactly one person (one absence downgrades the whole store). The "rostered every day without a day off" test is spelled out because knowing about a skill single point and rostering its only holder to exhaustion are two halves of the same mistake. Group 4 gives "close-only for months" its own line - the most common and least-complained-about fairness problem in stores. The people it touches usually don't argue; they quietly start looking for the next job.

Variables