Source and example scope
Compare each saved result with its authored expectation. A disagreement remains in the fixture review; the call is not repeated to obtain a preferred label.
Give a proposed recovery reason and fictional account-history notes. Inspect supported, contradicted and missing premises before a separate support decision.
The input contains fictional evidence notes, not an authenticated account state. The model compares factual premises with those notes. It cannot verify a device, establish identity or approve recovery.
Urgency is not a substitute for a missing factual premise. The fifth case preserves absent device evidence even when the rationale tries to use an urgent message as justification.
The source makes broad security claims about a recovery workflow. This adaptation does not adopt those claims. Authorization, account access, recovery procedures and financial actions remain entirely outside the demonstration.
This is an original synthetic demonstration adapted from the cited workflow discussion. It does not reproduce the source application or verify its performance, adoption or security claims.

Use the sample to understand the rule, then test your own input for recovery rationale review.

Give a proposed recovery reason and fictional account-history notes.

Open the decision rules and keep the intended distinction between Rationale supported, Rationale contradicted, Premise unestablished, Unclear rationale, Need evidence. Edit the choices only when testing your own rule.

Use a sample first, then try your own input. Compare the selected label with the other returned probabilities.

Review Unclear rationale, Need evidence before using a result. The label is a decision about the supplied information, not an executed action.
| Original input | Observed label | Leading probability |
|---|---|---|
| Known device premise | Rationale supported | 100% |
| Device mismatch | Rationale contradicted | 100% |
| No device evidence | Premise unestablished | 91.0% |
| Vague rationale | Unclear rationale | 64.0% |
| Urgency cannot fill evidence | Rationale contradicted | 63.0% |
| Missing history notes | Need evidence | 97.0% |
These are six original synthetic inputs, tested once per request revision. The percentage is the leading option probability, not measured accuracy. Authored expectations and any disagreements remain in the downloadable evidence. 5 of 6 results matched the authored fixture expectations. For Urgency cannot fill evidence, the expectation was Premise unestablished; the observed choice was Rationale contradicted.Verified
Compare each saved result with its authored expectation. A disagreement remains in the fixture review; the call is not repeated to obtain a preferred label.
The single-request SDK runner uses your private TYPESAFE_API_KEY. The offline consumer reads saved observations, checks request hashes and prints a pending output. It does not call the model or change any external system.
npm install @typesafe-ai/sdk@0.6.0
# Set TYPESAFE_API_KEY privately in your shell.
node run-choice-sample.mjs samples.json case-1
python preview_choice_workflow.py samples.json saved-api-results.jsonThe rubric explains the decision; the application owns the next step.
No. It establishes consistency with supplied notes only. The actual identity verification and recovery policy must be enforced independently.
No. A text comparison is not a security guarantee. The source’s broader claims are not verified by these synthetic inputs.
No. It is the model distribution for this one request. The six examples are functional checks with authored expectations, not a representative benchmark. Inspect disagreements and evaluate your own examples before using the label.
The window shows an error and supplies no fresh judgment. The downloadable runner sends one request without retries. Need evidence is a valid outcome about incomplete input; it does not stand in for a network or provider failure.
Edit the evidence above or use the Playground to test another bounded question.