Case material
ready20 converted source artifacts and 5 prompt steps are loaded for rehearsal.
blocked-before-live-rehearsal
Read-only operator runbook for the first approved live OpenAI smoke test. It does not provision credentials, approve gates, or execute the model.
20 converted source artifacts and 5 prompt steps are loaded for rehearsal.
Review saved case overrides before live execution; registry defaults remain versioned in git.
Run validation cases around the live smoke result to catch retrieval misses and source drift.
Human action: provision OPENAI_API_KEY for this app process, restart the server, then run live validation.
Record accepted/rejected local review decision for Data handling policy before live rehearsal.
Record accepted/rejected local review decision for Access and role model before live rehearsal.
Record accepted/rejected local review decision for Live OpenAI validation before live rehearsal.
Record accepted/rejected local review decision for Output and archive boundary before live rehearsal.
Record accepted/rejected local review decision for Anonymized learning before live rehearsal.
Confirm the current no-go/go wording before any live action.
/cases/CASE-DEMO-001/decision-centerResolve or explicitly defer blocker rows that affect live validation.
/cases/CASE-DEMO-001/action-boardConfirm case-scoped overrides and source variables are appropriate for the smoke run.
/cases/CASE-DEMO-001/promptsConfirm secret-handling rules, restart steps, and post-provisioning evidence before live execution.
/cases/CASE-DEMO-001/provisioningUse the readiness page button only after approvals and OPENAI_API_KEY are in place.
/cases/CASE-DEMO-001/readinessOpen the created run detail from the run list and verify every step is source-traceable.
/cases/CASE-DEMO-001/auditCapture the Markdown review pack after the smoke run for internal evidence.
/cases/CASE-DEMO-001/review-pack.mdScreenshot or JSON showing OPENAI_API_KEY present without exposing the secret.
/cases/CASE-DEMO-001/readinessApproval rows for data handling, access/role, live OpenAI, output/archive, and anonymized learning.
/cases/CASE-DEMO-001/approvalsStored run page with mode, selected sources, step outputs, and raw payload.
/cases/CASE-DEMO-001/runs/<run_id>Local records for readiness, run-detail, audit, decision, and review-pack evidence.
/cases/CASE-DEMO-001/live-evidenceMachine-readable go/no-go state after the smoke run.
/cases/CASE-DEMO-001/decision-center.jsonInternal Markdown handoff with source ids and model-produced run evidence.
/cases/CASE-DEMO-001/review-pack.mdDo not run live generation until the gate is recorded and accepted for the rehearsal scope.
/cases/CASE-DEMO-001/approvalsThe live route must fail loudly and should not be retried until the server is restarted with the key.
/cases/CASE-DEMO-001/readinessStop if a prompt requests unavailable or non-approved client material.
/cases/CASE-DEMO-001/promptsDo not use the output externally; add source evidence or revise the prompt before another run.
/cases/CASE-DEMO-001/canvasStop and confirm retention/access/legal handling before the material enters live processing.
/cases/CASE-DEMO-001/sourcesStructured runbook state for the smoke-test handoff.
/cases/CASE-DEMO-001/live-rehearsal.jsonCredential presence, restart, secret-handling, and evidence checklist.
/cases/CASE-DEMO-001/provisioning.jsonRecorded smoke-test evidence bundle and missing evidence rows.
/cases/CASE-DEMO-001/live-evidence.jsonOpen blockers and human action evidence.
/cases/CASE-DEMO-001/action-board.jsonFinal go/no-go language for internal review.
/cases/CASE-DEMO-001/decision-center.jsonDerived events around conversion, approval, run, feedback, and governance surfaces.
/cases/CASE-DEMO-001/audit-log.jsonHuman-readable source-traceable internal pack.
/cases/CASE-DEMO-001/review-pack.md