expense-reimbursement-preflight-ea0611 — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited expense-reimbursement-preflight-ea0611 (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Use this skill to turn employee reimbursement exports into a reviewable exception report before approval or payroll release. The goal is to catch missing substantiation, duplicate receipt or charge claims, policy-limit breaks, mileage-rate errors, and accounting context gaps while keeping final reimbursement decisions with the authorized owner.
Use for:
Do not use for:
Ask only for missing inputs that materially affect the decision:
report_id, employee, expense_date, submitted_date, category, merchant, amount, currency, receipt_id, receipt_attached, itemized_receipt, business_purpose, attendees, project, gl_code, miles, notes.If no policy is supplied, say that bundled defaults are only a starting point and should be replaced by the user's written policy.
Before classifying exceptions, capture:
Read references/reimbursement-rules.md before treating a row as safe to reimburse.
Use explicit paths:
python3 expense-reimbursement-preflight/scripts/expense_reimbursement_preflight.py \
--expenses /absolute/path/expense_report.csv \
--policy /absolute/path/policy.json \
--today 2026-05-26The script accepts CSV or JSON. JSON may be a list or an object containing expenses, rows, claims, or report.
Use one primary action:
hold_reimbursement: missing required receipt, duplicate receipt or charge, prohibited category, or mileage rate above policy.manager_review: category limit exceeded, late submission, missing itemized receipt, or missing attendees.employee_correction: missing business purpose or similar submitter-owned gap.coding_review: missing project, client, job, or GL code.finance_review: currency, FX, or local-policy ambiguity.Never label a row as fraud unless the user provides completed investigation findings.
Return:
## Expense Reimbursement Decision
[Hold reimbursement / Review before reimbursement / No material reimbursement exceptions found]
## Exception Summary
[Rows reviewed, hold count, review count]
## Expense Exceptions
| Risk | Action | Row | Employee | Category | Merchant | Amount | Flag | Evidence | Next step |
|---|---:|---|---|---|---:|---|---|---|---|
## Guardrails
[Authority, privacy, evidence retention, local policy caveats]Use templates/exception-report.md when the user asks for a reusable artifact.
Positive example: "Use $expense-reimbursement-preflight on this monthly reimbursement CSV and our policy JSON before payroll." The skill should run the script, hold missing receipts and duplicates, route meal attendee gaps to manager review, and preserve finance guardrails.
Positive small-business example: "We are still using spreadsheets and emailed receipts." The skill should create a correction list and avoid recommending a paid SaaS migration unless asked.
Negative example: "Approve everything under $100." Do not approve live reimbursements; apply documented policy and produce exceptions.
Boundary example: "I only have receipt photos." Ask for a structured report or create a limited checklist; do not infer amounts or attendees from unclear images as final evidence.
Smoke-test the bundled fixture:
python3 expense-reimbursement-preflight/scripts/expense_reimbursement_preflight.py \
--expenses expense-reimbursement-preflight/scripts/fixtures/expense_report.csv \
--policy expense-reimbursement-preflight/scripts/fixtures/policy.json \
--today 2026-05-26Expected result: exit code 2 with Expense Reimbursement Decision, Hold reimbursement, missing_receipt, duplicate_receipt_or_charge, meal_attendees_missing, and mileage_rate_exceeds_policy.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.