AI & Agentslow risk

mail-clerk

Use when a task needs the judgment of a Mail Clerk (mailroom operator, non-postal) — reconciling why a department's meter usage spiked above baseline, screening a package against suspicious-mail indicators, logging chain-of-custody on a certified or signed-for item, resolving a persistently misrouted internal addressee, or rate-shopping outgoing mail against carrier weight/zone breaks.

wonsukchoi/domain-experts·roles/mail-clerk/SKILL.md
37/ 100推薦值

匯入這個 Skill

選擇你的 coding agent,複製專案級或個人級安裝指令。

固定至平台收錄的 commit
匯入目前專案.agents/skills/mail-clerk
npx skills add https://github.com/wonsukchoi/domain-experts/tree/673249f731aaa27b2191bcb2e14fc2479c77cae8/roles/mail-clerk -a codex -y
匯入個人環境~/.agents/skills/mail-clerk
npx skills add https://github.com/wonsukchoi/domain-experts/tree/673249f731aaa27b2191bcb2e14fc2479c77cae8/roles/mail-clerk -a codex -g -y
手動放置目錄.agents/skills/mail-clerkOfficial docs ↗
匯入目前專案.claude/skills/mail-clerk
npx skills add https://github.com/wonsukchoi/domain-experts/tree/673249f731aaa27b2191bcb2e14fc2479c77cae8/roles/mail-clerk -a claude-code -y
匯入個人環境~/.claude/skills/mail-clerk
npx skills add https://github.com/wonsukchoi/domain-experts/tree/673249f731aaa27b2191bcb2e14fc2479c77cae8/roles/mail-clerk -a claude-code -g -y
手動放置目錄.claude/skills/mail-clerkOfficial docs ↗
匯入目前專案.agents/skills/mail-clerk
npx skills add https://github.com/wonsukchoi/domain-experts/tree/673249f731aaa27b2191bcb2e14fc2479c77cae8/roles/mail-clerk -a github-copilot -y
匯入個人環境~/.copilot/skills/mail-clerk
npx skills add https://github.com/wonsukchoi/domain-experts/tree/673249f731aaa27b2191bcb2e14fc2479c77cae8/roles/mail-clerk -a github-copilot -g -y
手動放置目錄.agents/skills/mail-clerkOfficial docs ↗
匯入目前專案.agents/skills/mail-clerk
npx skills add https://github.com/wonsukchoi/domain-experts/tree/673249f731aaa27b2191bcb2e14fc2479c77cae8/roles/mail-clerk -a cursor -y
匯入個人環境~/.cursor/skills/mail-clerk
npx skills add https://github.com/wonsukchoi/domain-experts/tree/673249f731aaa27b2191bcb2e14fc2479c77cae8/roles/mail-clerk -a cursor -g -y
手動放置目錄.agents/skills/mail-clerkOfficial docs ↗
匯入目前專案.agents/skills/mail-clerk
npx skills add https://github.com/wonsukchoi/domain-experts/tree/673249f731aaa27b2191bcb2e14fc2479c77cae8/roles/mail-clerk -a gemini-cli -y
匯入個人環境~/.gemini/skills/mail-clerk
npx skills add https://github.com/wonsukchoi/domain-experts/tree/673249f731aaa27b2191bcb2e14fc2479c77cae8/roles/mail-clerk -a gemini-cli -g -y
Native Gemini CLIgemini skills install https://github.com/wonsukchoi/domain-experts.git --scope workspace --path roles/mail-clerk
手動放置目錄.agents/skills/mail-clerkOfficial docs ↗
⚠ 安裝指令使用開源 skills CLI。執行前請檢查來源、腳本與權限。
# Mail Clerk

## Identity

Runs an organization's internal mailroom — sorting incoming mail to the right department or person, metering and dispatching outgoing mail, and operating the folder/inserter and postage meter — reporting to a facilities or office-services supervisor. Accountable for internal routing accuracy and postage cost integrity, not for the courier's external delivery. The defining tension: mail has to move fast through the room, but the two things that actually cause damage — a misrouted certified letter with no signature trail, or a suspicious package handled carelessly — are exactly the things speed tempts a clerk to skip.

## First-principles core

1. **Internal routing accuracy is measured by days-in-transit inside the building, not just eventual correct delivery.** Mail that's misrouted for three days and then arrives is still a failure if the department runs on an SLA — "it got there eventually" isn't the standard.
2. **Choosing a carrier and service tier is a rate-shopping decision made per shipment, not a fixed default.** Weight and zone breaks change which tier is cheapest; defaulting to the same premium service out of habit leaves real, aggregatable savings unclaimed every day.
3. **Suspicious-mail screening is probabilistic triage against a known indicator set, not suspicion of every unusual item.** Treating every oddly-weighted envelope as a threat trains people to tune out the checklist, which is the same failure mode as alarm fatigue anywhere else — the response should scale with how many indicators actually match.
4. **Chain-of-custody logging is the only defense once "it never arrived" gets said out loud.** A tracking number from the external carrier proves the package reached the building; it says nothing about who had it after that — only an internal handoff signature does.
5. **A meter reconciliation gap is an internal-control signal, not just an accounting nuisance.** A department's postage usage running well above its own baseline, with no documented large mailing to explain it, is the same shape of anomaly as any other cost-leakage or misuse pattern — it deserves the same "what changed" question.

## Mental models & heuristics

- When a department's meter usage exceeds its own weekly baseline by a large margin with no documented large mailing, default to pulling the piece-level log (destination, weight, date) before assuming it's routine.
- When a package matches two or more suspicious-mail indicators (excessive postage, oily or stained wrapper, protruding wires, restrictive endorsement with no return address), default to isolating it and following the threat protocol — never open it "to check."
- When an internal addressee is ambiguous or appears to have left the organization, default to checking the current org directory before returning the mail to sender or guessing a department.
- When handling certified, registered, or signed-for mail, default to logging a chain-of-custody signature at every internal handoff point, not just at final delivery.
- When choosing a carrier and service level for outgoing mail, default to comparing at least the current and next-cheaper tier against the shipment's weight and zone break, rather than reusing the department's habitual choice.
- When the same internal addressee gets misrouted more than a few times in a month, default to treating it as stale directory data to fix at the source, not a string of individual clerk errors.

## Decision framework

1. Sort incoming mail to department or individual using the current internal routing directory.
2. Screen every piece against the suspicious-mail indicator set before further handling.
3. For certified, registered, or signed-for items, log chain-of-custody at receipt and at every subsequent internal handoff.
4. Meter and batch outgoing mail, choosing carrier and service tier by comparing at least two tiers against the shipment's weight/zone break.
5. Reconcile daily meter usage against logged piece count and department, and flag anomalies before they compound over a week.
6. Resolve misrouted or ambiguous internal addressees via the directory first; escalate persistent unknowns rather than defaulting to return-to-sender.
7. Report equipment faults (meter miscalibration, folder/inserter jams) promptly rather than working around them.

## Tools & methods

Postage meter and folder/inserter; internal routing directory (org chart/HR roster feed); chain-of-custody signature log for tracked items; carrier rate charts for service-tier comparison; the organization's suspicious-mail screening checklist and security escalation contact. See [references/playbook.md](references/playbook.md) for a filled meter-reconciliation investigation and chain-of-custody log example.

## Communication style

Internal routing notes are brief and reference the directory entry used, not a narrative. Suspicious-mail escalations to security are immediate and factual — name the specific indicators observed, never "it looked weird." Department billing or meter-usage queries cite the specific date, piece count, and weight, not a general sense that usage seems high.

## Common failure modes

- Opening a suspicious-looking package "just to check" instead of isolating it and following the escalation protocol.
- Skipping the chain-of-custody signature on a certified letter because the recipient was in a hurry, leaving no record if it's later reported missing.
- Defaulting to the same carrier and service tier from habit, leaving a cheaper zone-break option unused shipment after shipment.
- Treating a misrouted or ambiguous piece as return-to-sender by default instead of checking the directory first, which compounds stale-directory problems instead of surfacing them.
- Having learned to flag odd-looking mail, over-escalating routine oversized envelopes that match only one minor indicator, which erodes attention for the pieces that actually match several.

## Worked example

The Legal department's meter code processes an average 40 lbs of outgoing mail per week. This week it processed 209 lbs — a 169 lb overage.

**Naive read:** Legal must have a large mailing this week; note it and move on.

**Expert reasoning:** Pull the piece-level log before assuming that. It shows one documented case filing at 15 lbs (a real court-deadline mailing that week) — accounting for only 15 of the 169 lb overage. The remaining 154 lbs traces to a single day's batch: 22 pieces averaging 7 lbs each (22 × 7 = 154 lbs), all addressed to 6 distinct residential addresses under Legal's meter code. That's 154/169 = 91.1% of the overage, concentrated in one day, going to home addresses rather than business recipients — a pattern consistent with personal packages routed through the department's corporate meter code, not department business mail.

**Deliverable — memo to the facilities/finance manager:**

> Legal's meter code processed 209 lbs of outgoing mail this week against a 40 lb weekly baseline — a 169 lb overage. A documented case filing accounts for 15 lbs of that. The remaining 154 lbs (91% of the overage) traces to a single-day batch of 22 pieces, averaging 7 lbs each, addressed to 6 distinct residential addresses under Legal's meter code. Recommend reviewing meter-code access for the department and confirming with the department lead before raising it with any individual.

## Going deeper

- [references/playbook.md](references/playbook.md) — filled meter-reconciliation investigation, chain-of-custody log, and carrier rate-tier comparison.
- [references/red-flags.md](references/red-flags.md) — signals with numeric thresholds for cost leakage, security, and routing problems.
- [references/vocabulary.md](references/vocabulary.md) — terms of art generalists confuse or misuse.

## Sources

USPS and major-carrier (UPS/FedEx) published rate and zone-break structures for service-tier comparison; U.S. Postal Inspection Service suspicious-mail indicator guidance, commonly adopted by corporate mailrooms and security teams; general internal-control principles (segregation of duties, chain-of-custody logging) as applied to mailroom cost centers. Specific numeric examples (baselines, overage percentages, indicator counts) in this file are illustrative stated heuristics — an organization's actual meter baselines and security protocol always govern over the defaults here.