ILLUSTRATIVE EXAMPLE — SYNTHETIC RECORD
“Warning appeared intermittently.”
No exact code, operating context, or source timestamp is supplied.
TRODEON · Decision support in development
Trodeon is being developed to compare maintenance signals with candidate causes and make supporting, conflicting, stale, and missing information easier to review. Any future predictive or diagnostic output would be a hypothesis, not a confirmed fault; fleet personnel choose what to inspect, repair, or dispatch.
No live vehicle feed, provider integration, independently validated performance, or customer-fleet outcome is established. The local software checks below do not validate diagnosis or prediction on a customer vehicle.
How the review stays grounded
Explore the synthetic chronology, see the missing and unconfirmed evidence, and choose a local-only human review question. No choice starts an operational action.
TRODEON · PROPOSED REVIEW FORMAT
ILLUSTRATIVE EXAMPLE — SYNTHETIC RECORDS
This synthetic walkthrough brings observations, previous work, uncertainty, and missing context into view for a fleet manager. It demonstrates a review format, not a diagnosis or prediction.
Not live vehicle data. No customer case, validated fault, repair instruction, service dispatch, or operational action is represented.
01 / RECORD CHRONOLOGY
ILLUSTRATIVE EXAMPLE — SYNTHETIC RECORD
“Warning appeared intermittently.”
No exact code, operating context, or source timestamp is supplied.
ILLUSTRATIVE EXAMPLE — SYNTHETIC RECORD
“Inspected reported concern.”
Inspection findings, repair details, and follow-up outcome are missing.
ILLUSTRATIVE EXAMPLE — SYNTHETIC RECORD
“Warning reported again.”
The available notes do not establish whether this matches the earlier warning.
A prior selected local Vitest run recorded on October 12, 2026 passed 14 files and 68 tests. The database-backed and HTTP acceptance results below are earlier local records documented in docs/fleet-reviewer-isolation-tech-notes.md. This public-site redesign did not rerun those suites; none of those checks establishes diagnostic accuracy, predictive quality, staging acceptance, or live-fleet outcomes.
| Evidence area | Conditions and assertions | Source references | Verification and limits |
|---|---|---|---|
| Evidence validation and abstention | Local unit tests use checked-in synthetic cases for missing, stale, duplicate, conflicting, unmapped, sufficient, and future-time evidence. Assertions check missing-field reasons, evidence findings, and an explicit ABSTAIN decision when the evidence does not establish a diagnosis or prediction. | tests/unit/business/diagnostic-evidence-review.test.ts; tests/unit/contracts/diagnostic-evidence-review.test.ts | A selected local Vitest run recorded on October 12, 2026 passed 14 files and 68 tests. The two cited evidence-review tests use checked-in synthetic fixtures; their assertions are limited to the listed cases. Synthetic tests do not establish diagnostic accuracy, prediction quality, or behavior on a customer vehicle. Diagnosis and prediction remain NOT_ESTABLISHED. |
| Onboarding authorization gates | Local business and route tests check missing or mismatched approvals, stopped partial intake, server-side authorization decisions, and rejection of unsafe scope changes. The route suite uses mocked session and storage responses. | tests/unit/business/fleet-evaluation-onboarding.test.ts; tests/unit/integrations/fleet-evaluation-onboarding-route.test.ts | A selected local Vitest run recorded on October 12, 2026 passed 14 files and 68 tests. The named route tests use mocked session and storage dependencies. These checks do not replace a real browser-session review or staged acceptance. |
| Review-record persistence | Unit route tests exercise scoped create, read, and same-scope replay while rejecting altered scope, foreign record identifiers, client-supplied identity or evidence, and unsupported alter/export operations. | tests/unit/integrations/diagnostic-evidence-review-route.test.ts | A selected local Vitest run recorded on October 12, 2026 passed 14 files and 68 tests. This route test mocks the database adapter and session; it is not a database-backed persistence test. Mocked persistence confirms route behavior under the mock only. Database and staged acceptance remain separate. |
| Fleet-scoped access | A separate database-backed acceptance test creates unique synthetic users, records, and two fleet scopes; it checks foreign identifiers, altered scope parameters, same-scope review behavior, cross-scope conflict, onboarding scope, revocation, and a vehicle-link collision. | tests/unit/integrations/fleet-reviewer-isolation-db.test.ts | A prior disposable-database run is recorded in docs/fleet-reviewer-isolation-tech-notes.md as 1 file and 3 tests passed on 2026-10-12. It was not rerun for this public-site redesign. That recorded run used synthetic rows, actual route handlers and Prisma, mocked authentication, and blocked provider fetches. Authentication was mocked. This does not replace browser-session checks, staged acceptance, live fleet verification, or production security review. |
| Fleet reviewer isolation · local acceptance | A local acceptance run exercised real Better Auth email/password sign-in and session retrieval; authorized fleet-scoped review save/reload; denial of a foreign review id, changed tenant/fleet, and another reviewer; cross-origin POST rejection without a persisted review; and review plus session persistence after a safe local app restart. | scripts/fleet-reviewer-isolation-live-check.mjs; /api/diagnostic-evidence-review-scopes; /api/diagnostic-evidence-reviews | A prior local HTTP run is recorded in docs/fleet-reviewer-isolation-tech-notes.md as 11 named checks passed on 2026-10-12. It was not rerun for this public-site redesign. The recorded run used real local Better Auth sign-in/session retrieval and Prisma with synthetic reviewers and records in a disposable local database; endpoints were localhost only. Local verification only. The browser test was skipped and no fleet-isolation screenshots are available. Staging acceptance and live-fleet validation remain pending; no production or fleet outcome is established. |
| Integrity and tamper checks | Local unit tests check stable canonicalization and signing of synthetic review records, rejection after content, scope, MAC, or secret changes, and review-packet malformed, digest-mismatch, and broken-chain outcomes. | tests/unit/business/diagnostic-evidence-review-record.test.ts; tests/unit/business/diagnostic-evidence-review.test.ts | A selected local Vitest run recorded on October 12, 2026 passed 14 files and 68 tests. Test-only signing values are fixtures; no production secret is used. These specific checks do not establish immutability, tamper-proofing, independent authentication, or production security. |
What remains
Browser UI verification, staging acceptance, pilot-specific provider permissions and available fields, and outcomes from a controlled fleet pilot remain to be established. The local acceptance run does not validate a live fleet.
These checks establish specific software behaviors under test conditions. They do not establish reduced downtime, avoided breakdowns, diagnostic accuracy, broad vehicle compatibility, or production security.
Interested in the 90-Day Downtime Prevention Audit?
Review the Fleet Evaluation