TaxFlo / SafeVault field note - prepared October 20, 2026
Fictional exercise. All eight cases NOT RUN. Test an existing workflow with these five fictional text sources. This is a proposed evaluation method, not a product certification, IRS form, retention policy or tax judgment.
Start with files 01 and 02, then introduce 03 and 04 in sequence. File 05 tests useful historical reuse. Use only authorized demonstration roles and environments. No taxpayer documents.
| Case | Question | Setup | Expected observation | Failure boundary | Actual record |
|---|---|---|---|---|---|
| PY1 | Prior statement only | Load PY-01 and ORG-01. Leave current support absent. | Identify 2025 context separately; current account status/support remains unresolved. | Fail if retained evidence automatically closes the 2026 question. | NOT RUN Observed result: __________ Evidence: __________ Owner / next action: __________ |
| PY2 | Blank answer survives carry-forward | Carry forward the historical account reference, leaving ORG-01 blank. | The current response remains blank/unresolved; assign a next owner. | Fail if an old answer is presented as current without explicit confirmation. | NOT RUN Observed result: __________ Evidence: __________ Owner / next action: __________ |
| PY3 | Current support arrives | Introduce CUR-01 version 1. | Identify its year, version and inspection route; practitioner decides sufficiency. | Do not equate receipt with final tax approval. | NOT RUN Observed result: __________ Evidence: __________ Owner / next action: __________ |
| PY4 | Correction after review | Have reviewer A record a fictional readiness decision against version 1; then introduce version 2. | Retain which version supported the earlier decision and route the correction for impact review. | Fail if two versions are counted as independent corroboration or earlier review silently transfers. | NOT RUN Observed result: __________ Evidence: __________ Owner / next action: __________ |
| PY5 | Handoff before resolution | Move the open correction question from reviewer A to reviewer B through approved controls. | Name current owner, remaining question and acceptance of next action; inspect old decision context. | Fail if reassignment erases the unresolved question. | NOT RUN Observed result: __________ Evidence: __________ Owner / next action: __________ |
| PY6 | Historical context remains useful | Add HIST-02 and ask which source provides acquisition context. | Locate the retained source with period and purpose; leave tax adequacy to practitioner. | Fail if all historical evidence is rejected solely because it is old. | NOT RUN Observed result: __________ Evidence: __________ Owner / next action: __________ |
| PY7 | Authorized reuse | Use only a preauthorized demonstration environment; compare intended authorized and unauthorized test roles under approved policy. | Record actual observed access and approved policy expectation separately. | Do not change permissions or claim a control passed without authorized execution. | NOT RUN Observed result: __________ Evidence: __________ Owner / next action: __________ |
| PY8 | Reproducible handoff record | Ask another reviewer to reconstruct the final state from the saved record. | Identify retained context, current source/version, unresolved change, owner and next action. | Do not invent timing, improved accuracy or a completed pilot outcome. | NOT RUN Observed result: __________ Evidence: __________ Owner / next action: __________ |
For one baseline and one comparison, record the same reviewer task, source set, start/stop rule, time and unresolved-error count. Leave blank until measured. A fictional expected observation is not proof of implemented TaxFlo behavior or a customer outcome. The practitioner decides evidence adequacy, tax treatment and final approval.
Framework and fixtures by Raman Yakkala, founder of AmitaSoft, building TaxFlo on SafeVault. IRS Recordkeeping explains purposes of records and event-dependent retention; checked October 20, 2026. Publication 4557 informs the article's access-control boundary. These sources do not prescribe this matrix or certify the product.
Stage companion article: Prior-year tax documents: reuse evidence, recheck assumptions. Stage preview only. The Friday article and this reader attachment remain unpublished to Production pending release review.