Local review and promotion loop
How a generated Hyphenomenon project stays editable and accountable after the first intake run: local pages expose the draft record, the operator reviews narrative, artifacts, workflows, and graph links, then decides whether to revise the packet, keep the record local, or later request a separate production apply.
Workflow steps
- 1
Record a checkpoint with absolute timestamp, dry-run/apply artifacts, verification, and next review decision
Action
- 2
Check whether the record explains what happened, what artifacts prove it, what AI did, and what the human operator decided
Process
- 3
Open the local project route, workflow route, notes, and Explore map created by the intake apply
Action
- 4
Keep production apply separate until explicitly approved with a fresh dry-run and scoped mutation plan
State
- 5
Decide whether the output is a useful proof record or whether the packet/dossier needs another rewrite
Decision
- 6
If weak, revise packet copy or workflow logic and rerun dry-run before another local apply
Action