sibyl-human-alignment

To align product intent with execution tracking, you can run the 'sibyl align' command. This refreshes the human-alignment store by serializing artifacts and regenerating architecture models, divergence reports, and decision ledgers.

A human-alignment store is a directory containing architecture models, divergence reports, decision ledgers, and narrative exports. It ensures that agentic engineering execution remains continuously aligned with the original product intent.

Yes, an isolated subagent is required to generate architecture diagrams and narratives. The alignment process delegates generative narrative and diagram authoring tasks to this subagent while the harness CLI handles artifact serialization.

You should regenerate divergence reports and decision ledgers at each story wrap or reflect cycle. This lifecycle integration ensures the Human Alignment cockpit tab always reflects the most current execution state and product intent.

No, the SIBYL harness CLI requires an isolated subagent to function correctly for this task. While the CLI handles artifact serialization, the subagent is strictly required for the generative tasks of authoring narratives and diagrams.