# Procurement Policy Check

> Decide whether one proposed procurement request is allowed under an explicit procurement policy before any spend flow starts, refusing budget overages and unapproved payees instead of guessing.

- Skill: `runxhq/procurement-policy-check` (Agent Skill, multi-file: 10 files)
- Install (CLI): `npx get-skill add runxhq/procurement-policy-check`
- Raw SKILL.md: https://api.skillmd.com/api/skills/runxhq/procurement-policy-check/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Finance & Business
- Author: runxhq (https://skillmd.com/u/runxhq)
- Updated: 2026-11-11
- Page: https://skillmd.com/skills/runxhq/procurement-policy-check

---


# Procurement Policy Check

Procurement Policy Check is the pre-spend decision gate. It answers one
question, is this procurement request allowed under the supplied procurement
policy, and it answers it deterministically. It never moves money; `pay` and
the payment rails own execution, and this skill's sealed decision is what a
caller carries into that flow.

## Procedure

1. Native `data.digest` binds the exact procurement request and procurement
   policy before any judgment.
2. Deterministic rules apply in order: the payee must be in
   `approved_payees`, the request currency must match the policy currency, and
   the amount must fit both `remaining_budget` and `request_cap`. The
   first failed rule refuses with a named `refused_reason`.
3. Zero remaining budget/cap and an empty payee list are explicit deny settings,
   not missing evidence. A zero-priced procurement is valid; no payment is implied.
   An in-policy amount that reaches `approval_threshold` returns `needs_human`
   with the policy's `human_lane` (default `procurement-review`). A zero threshold
   requires review for every request. Below the threshold the policy result is
   `approved`, which is not a native grant or approval to spend.
4. A policy missing any required authority field refuses with
   `missing_policy_authority` and findings; nothing is inferred or invented.

`spend_executed` is always false. The decision packet carries both input
digests so a downstream spend can prove which request and policy were decided.

## Output

`procurement_approval` (`runx/procurement-policy-check.approval.v1`) carries `decision`
(`approved`, `refused`, `needs_human`), `reason`, `payee`, `amount`,
`currency`, `ceiling_amount`, `human_lane`, `refused_reason`, `policy_ids`,
`validation`, and the two digests.

Inputs are `procurement_request` and `procurement_policy`.

## Evidence and execution boundary

The host selects the current authoritative policy and available-budget snapshot.
All compared amounts must use the same declared currency denomination; this
skill performs no unit, FX or provider conversion. Retain the original source,
observation time and both native digests with the decision.

A snapshot is not a reservation. Concurrent decisions can each fit the same
remaining budget; the selected payment/execution owner must refresh the relevant
policy and enforce its own atomic budget, exact amount, authority and native
approval before spending. Include any procurement review requirement in that
owner's approval context, not a second caller-authored approval flag.
This result never records payment, reserves funds or prevents another buyer
from using the budget. Use it as a deterministic policy check, not a ledger.

