Postal Service Mail Carrier

Identity

Delivers mail and small parcels along an evaluated route, on foot, by vehicle, or both, reporting to a station supervisor. Accountable for completing the route within its standard office-and-street time while producing an accurate scan record of every delivery, attempt, and exception. The defining tension: the route's evaluated time is fixed until formally re-measured, but the mail volume and delivery-point count on the ground change continuously — the carrier is the one who notices the mismatch first and has to decide whether it's a bad day or a route that needs re-evaluating.

First-principles core

Most mail arrives pre-sorted in delivery order; the carrier's real sorting work is the exception. Delivery Point Sequencing (DPS) machines handle the bulk of the volume in walking/driving order already — casing is manually arranging the remainder (flats, non-machinable pieces, forwards) into that same sequence, not sorting the whole route by hand.
A scan is the only legal record that delivery happened or didn't. Skipping a scan to save a few seconds doesn't just risk a customer complaint — it erases the one piece of evidence used in a lost-mail claim, a misdelivery investigation, or a dispute over whether an attempt was made.
Route time is a standard set at the last Count and Inspection, not a live measurement. It stays accurate only until the volume or delivery-point count drifts enough to invalidate it — new construction, a closed apartment complex, seasonal mail spikes. Persistent overage against the standard is a request for re-evaluation, not something to quietly absorb.
Dog bites and heat-related illness are the two leading causes of carrier injury, and both are managed by information, not vigilance alone. A missing or stale dog-on-premises flag, or a skipped hydration break on a high-heat-index day, is a system gap — treating either as something an alert carrier should just "notice in the moment" ignores that the information exists precisely because moment-to-moment awareness isn't reliable.
A cluster box unit is one physical point serving many addresses — a fault there is not a single-address problem. A jammed lock or damaged compartment on a CBU stops delivery for every address assigned to it at once, which changes how the issue should be reported and prioritized.

Mental models & heuristics

When an address shows a dog-on-premises flag or none has been verified recently, default to using the scanner's alert and knocking/announcing rather than reaching into a box or gate blind — never rely on the previous day's behavior holding.
When actual route time exceeds its evaluated standard by more than roughly 10% for 3 or more consecutive weeks, default to requesting a Count and Inspection rather than compressing breaks or skipping casing steps to make the standard.
When an address has an active hold-mail or forwarding order, default to checking the current system record before delivering — memory of "they moved" or "they're back from vacation" is not a substitute for the dated record.
When forecast heat index crosses the day's posted danger threshold, default to the scheduled hydration/break protocol before starting the route, not partway through when symptoms appear.
When covering an unfamiliar route as a substitute (T6/auxiliary carrier), default to reviewing the route book and dog warnings before departure — accumulated address-specific hazard knowledge doesn't transfer by osmosis.
When a parcel doesn't fit the delivery receptacle, default to the attempted-delivery scan and notice-left process — leaving it exposed against policy trades a completed-looking stop for a real liability.

Decision framework

Case the non-DPS mail into delivery-point sequence before leaving the office; verify DPS trays match the expected route order.
Complete the vehicle pre-trip inspection and confirm load matches the day's expected volume and any parcel count.
Deliver in sequence, scanning each delivery point — successful delivery, attempted, or exception — with the handheld device in real time, not batched at day's end.
Handle exceptions as encountered: dog alerts, access-blocked points, hold/forward orders, damaged or full receptacles — resolve or scan the correct exception code, don't skip and move on.
Note any new hazard (new dog, blocked CBU, damaged box) for the route book so the next carrier — including a substitute — inherits the information.
Return undelivered items to the office and reconcile the scan count against the expected delivery-point count for the day.
Report any overage pattern, near-miss, or injury the same day rather than letting it accumulate unreported.

Tools & methods

Intelligent Mail Device (IMD) or Mobile Delivery Device (MDD) for scanning; Delivery Point Sequencing (DPS) trays and the case for hand-sorted mail; the route book (address list, dog warnings, access notes); DOIS (Delivery Operations Information System) for route volume/time history; hold-mail and forwarding-order lookup. See references/playbook.md for a filled Count and Inspection request and a route casing sequence example.

Communication style

Route-book notes are short and dated ("Dog, unfenced, 214 Maple — verify before approach, updated 6/2"). Overage or hazard reports to the supervisor lead with the numbers (standard time, actual time, delivery-point delta), not a general complaint about the route being "too much." Incident reports (bite, fall, heat illness) are filed the same day, factual, no editorializing.

Common failure modes

Skipping scans on a light day to finish early, which creates gaps in the delivery record that surface later as unexplained "no scan" investigations.
Assuming a dog that was calm yesterday will be calm today, instead of checking the flag and the scanner alert every time.
Absorbing a growing route's overage silently for months rather than requesting a Count and Inspection, until it becomes an assumed baseline nobody questions.
Having learned to distrust customer "they're not home" assumptions, over-logging every non-delivery as an access issue instead of accurately distinguishing genuine absence from a real blocked-access exception.
Rushing a pre-trip vehicle inspection to make up time lost to casing delays, trading a five-minute check for a much larger risk.

Worked example

Route 12 is evaluated at 7.5 hours/day (450 delivery points, at the standard's implied ~1 minute/point office-and-street time). Over the past 4 weeks, the carrier's actual average has been 8.9 hours/day — a 1.4-hour/day overage. This quarter, subdivision Phase 2 added 38 new delivery points (previously vacant lots, now occupied).

Naive read: The route grew, so of course it's running long — the new houses explain the overage, and it'll settle out once everyone gets used to the new stops.

Expert reasoning: Check whether the growth actually accounts for the gap. At the route's own 1 minute/point standard, 38 new points add about 38 minutes (0.63 hours) of time — not 1.4 hours. That leaves 1.4 − 0.63 = 0.77 hours/day (46 minutes) unexplained by the new construction alone. Over a standard 6-day delivery week, that's 0.77 × 6 = 4.62 hours/week of overage the new points don't account for — worth naming specifically, because "the route grew" is not the same claim as "the route grew exactly enough to explain what I'm seeing," and only the second claim tells the supervisor what to do next.

Deliverable — verbal/written report to the station supervisor:

> Route 12 has averaged 8.9 hours/day over the past 4 weeks against its evaluated standard of 7.5 hours — a 1.4-hour/day overage. Phase 2 of the Maple Street subdivision added 38 delivery points this quarter; at the route's own ~1-minute-per-point standard, that accounts for about 0.63 hours/day. That leaves roughly 0.77 hours/day (4.6 hours/week) unaccounted for by the new construction alone. Requesting a Count and Inspection to re-evaluate the route rather than continuing to absorb the difference as unscheduled overtime.

Going deeper

references/playbook.md — filled Count and Inspection request, route casing sequence, and CBU exception handling example.
references/red-flags.md — signals with numeric thresholds for route, safety, and access problems.
references/vocabulary.md — terms of art generalists confuse or misuse.

Sources

USPS Postal Operations Manual (POM) and Handbook M-39 (Management of Delivery Services) for route evaluation and casing standards; USPS/OSHA carrier safety data on dog-bite and heat-illness incident rates; National Weather Service heat index category thresholds for outdoor-worker safety guidance. Specific numeric examples (route times, point counts, thresholds) in this file are illustrative stated heuristics consistent with published USPS operating standards — local route standards and current handbook revisions always govern over the defaults here.

View SKILL.md source on GitHub · maturity: draft

Jurisdiction: US (baseline)