entity-tracker

Tracking manuscript continuity requires a single ENTITY_STATE.yaml file that records characters, locations, and organizations from foundation.md and chapters. This canonical YAML serves as the source of truth so downstream tools reference it instead of re-reading full chapters.

Manuscript memory is maintained by constructing and updating a single ENTITY_STATE.yaml file that aggregates entity data from foundation.md and chapters. This YAML format tracks relationships, timelines, and world rules to provide canonical data for downstream skills.

Detecting knowledge conflicts involves flagging contradictions as conflicts and surfacing knowledge gaps during continuity review. The entity-tracker cross-references entity data across chapters to identify inconsistencies and ensure every data point includes a source in ch-XX:pYY format.

Updating entity data in a YAML file supports both BUILD and UPDATE modes for adding new chapters. You can run the tracker to generate or update ENTITY_STATE.yaml from foundation.md and all chapters, tracking new and existing entities without losing prior canonical data.

Manuscript memory tracking operates with no external dependencies, relying solely on foundation.md and chapter files as input. It processes these sources to build and maintain the ENTITY_STATE.yaml file independently of other tools or frameworks.

Continuity audits need source citations because every data point must include a source in ch-XX:pYY format to verify provenance. This requirement ensures that entity states can be traced back to specific chapter locations for accurate conflict resolution and knowledge gap surfacing.