---
title: Testing Expansion Backlog
tags: [testing, backlog, qa, release]
created: 2026-06-06
---

# Testing Expansion Backlog

This is the testing backlog to complete before expanding the escrow test suite
or claiming release confidence.

## P0 - Product Policy Gaps

### Physical product grace period

Need implementation and tests for:

- seller marks delivery;
- buyer receives delivery window;
- grace period starts;
- buyer can confirm delivery before expiry;
- buyer can dispute before expiry;
- funds release only after buyer confirmation, no dispute, or grace expiry.

Acceptance tests:

| ID | Expected |
|---|---|
| RELEASE-PHYS-001 | Physical delivery starts hold/grace timer. |
| RELEASE-PHYS-002 | Buyer confirmation before expiry makes funds releasable. |
| RELEASE-PHYS-003 | Dispute before expiry blocks release. |
| RELEASE-PHYS-004 | Grace expiry without dispute makes funds releasable. |

### Immediate release for gift cards/digital goods

Need implementation and tests for:

- product type is gift card or digital;
- seller proof is required immediately;
- buyer acceptance or product policy triggers immediate/near-immediate release;
- dispute path can still stop release if policy includes a short challenge window.

Acceptance tests:

| ID | Expected |
|---|---|
| RELEASE-DIGI-001 | Gift card delivery does not use physical-product grace period. |
| RELEASE-DIGI-002 | Gift card proof advances to releasable immediately after buyer acceptance. |
| RELEASE-DIGI-003 | Wrong/missing proof blocks release. |

## P0 - Dispute Coverage

The dispute mechanism is not complete enough for end-to-end fund movement.

Tests needed:

| ID | Scenario | Expected |
|---|---|---|
| DISPUTE-001 | Buyer opens dispute after seller delivery, before release | payment/request enters dispute hold |
| DISPUTE-002 | Seller cannot force release while dispute is open | release attempt rejected |
| DISPUTE-003 | Admin resolves for seller | release becomes available |
| DISPUTE-004 | Admin resolves for buyer | refund becomes available |
| DISPUTE-005 | Duplicate dispute resolution | idempotent or rejected safely |
| DISPUTE-006 | Dashboard dispute route vs release-hold route | no route shadowing ambiguity |

## P0 - Payment Negative Tests

**Automated as of 2026-06-08** (`backend/__tests__/payment-edge-cases.test.ts`, 38 tests):
See [[Payment Safety Edge Cases]] for full detail.

- ✅ wrong token (on-chain `wrong_token` + direct-balance `address-token-mismatch` + `payment-wrong-token` event)
- ✅ wrong chain (direct-balance `address-token-mismatch`)
- ✅ wrong destination (direct-balance `address-token-mismatch`)
- ✅ underpayment (`insufficient_amount` on-chain; `underpaid` direct-balance + `payment-underpaid` event)
- ✅ native coin sent instead of ERC-20 (`wrong_asset` on-chain; stays pending in direct-balance webhook)
- ✅ OFAC-sanctioned sender blocked (opt-in per seller; direct-balance `fundDirectBalancePayment` AML gate)
- ✅ smart-contract sender blocked via `TRANSACTION_SAFETY_REQUIRE_EOA_SENDER=1`

Still needs automation:
- duplicate payment (double-credit guard);
- late payment after cancelled/expired intent;
- payment with no gas;
- scanner unavailable during payment;
- scanner webhook signature invalid (partially covered by `amn-pay-adapter-webhook-signature.test.ts`);
- balance check baseline missing or stale.

## P0 - Authorization and ID Boundaries

Add regression tests for cross-store id comparisons:

| Route/action | Required test |
|---|---|
| buyer delivery confirmation | buyer can confirm even when session id and request buyer id use different stores |
| buyer delivery confirmation | non-buyer cannot confirm |
| seller delivery update | only selected seller/admin can deliver |
| offer acceptance | only buyer/admin can accept |
| payment check | only buyer/admin/system can advance payment where applicable |

## P1 - UI/Browser E2E

Add Playwright tests for:

- checkout renders BSC Testnet chain label;
- token contract row shows canonical tUSDT;
- testnet explorer links are generated correctly;
- wrong wallet chain shows switch button;
- insufficient balance warning appears;
- payment submitted screen waits for scanner confirmation;
- paid socket/poll fallback transitions to success.

## P1 - Notification Assertions

Initial reusable E2E helpers are implemented in
`backend/scripts/smoke/marketplace-e2e-notifications.mjs`. Remaining work:

- capture per-user notification baseline;
- poll `GET /api/notifications` and `GET /api/notifications/unread-count`;
- optionally attach Socket.IO clients for buyer/seller/admin actors;
- fail on wrong-recipient notifications;
- record known notification gaps with route/action and expected recipient;
- include notification latency in every E2E report.

Current dev findings from the first two-round runner report:

- `NOTIF-E2E-001` passes for buyer new-offer notifications.
- Seller new-request fanout passes for privately targeted sellers.
- Rejected-seller notifications pass by rejected offer `relatedId`.
- `NOTIF-E2E-002` fails by selected offer `relatedId`; accepted-offer notification is not discoverable with the selected offer id.

Acceptance tests:

| ID | Expected |
|---|---|
| NOTIF-E2E-001 | New seller offer creates a buyer notification and unread-count increment. |
| NOTIF-E2E-002 | Buyer offer acceptance creates selected seller notification. |
| NOTIF-E2E-003 | Seller delivery creates buyer notification. |
| NOTIF-E2E-004 | Buyer delivery confirmation creates seller notification. |
| NOTIF-E2E-005 | Missing `pending_payment` notification is reported as a known gap, not a silent pass. |

## P1 - Concurrency and Performance Profile

Initial executable runner and report generator are implemented in
`backend/scripts/smoke/marketplace-e2e-notifications.mjs`. Remaining work:

- worker abstraction for one isolated buyer/sellers/payment/delivery flow;
- barrier start for simultaneous workers;
- ramp stages `1, 2, 4, 8, 16, 32`;
- JSON result per worker;
- markdown performance profile report per run;
- infrastructure snapshots before/during/after each stage;
- stop-condition enforcement.

Current dev blocker:

- Dev API rate limiting is `100` requests per `900s`, so C2+ ramp runs need a
  fresh window, a lower-polling profile, or a non-rate-limited test environment.
- `PAYMENT_MODE=record` currently fails because `POST /api/marketplace/payments`
  returns HTTP 500 after offer selection with generated test actors.

Acceptance tests:

| ID | Expected |
|---|---|
| PERF-CONC-001 | C1 completes with all notification assertions. |
| PERF-CONC-002 | C2/C4 detect no cross-worker state leakage. |
| PERF-CONC-003 | Report includes p50/p95/p99 API, scanner, notification, and total timings. |
| PERF-CONC-004 | Runner stops the ramp on P0 notification/payment/ledger correctness failure. |
| PERF-CONC-005 | Runner separates live-chain mode from scanner-fixture/API-only modes. |

## P1 - CI and Deployment

Add a deployment smoke checklist artifact after every pipeline:

- backend version endpoint;
- frontend root 200;
- scanner `/health`;
- chain 97 token balance check;
- current Docker image digests;
- last commit SHA per service.

## P1 - Test Data Lifecycle

Need cleanup strategy:

- flag generated users with run id;
- archive or tag generated purchase requests;
- prevent test data from polluting analytics;
- define whether test tx/payment rows are permanent audit evidence or can be purged.

## P2 - Observability

Add dashboards/log queries for:

- scanner intent lifecycle;
- balance-check deltas;
- payment state transitions;
- delivery status changes;
- dispute holds;
- release/refund attempts.

## Suggested Next Automation Order

1. Fix selected-seller accepted notification related id.
2. Fix or retire the legacy `POST /api/marketplace/payments` record path.
3. Run `PAYMENT_MODE=status` through delivery after the dev rate-limit window resets.
4. Run `PAYMENT_MODE=live` once wallet gas/tUSDT and rate-limit headroom are available.
5. Add non-buyer delivery confirmation regression test.
6. Add scanner negative tests for wrong token and underpayment.
7. Implement physical vs digital release policy.
8. Add release-policy tests.
9. Resolve dispute route/policy gaps.
10. Add dispute-to-release/refund E2E tests.