Corpus Reconcile

Synthesize raw project files into canonical documentation by running the corpus reconcile skill, which automatically clusters related files and uses subagents to draft structured project records. It transitions staged brainstorming notes into finalized assets within a defined repository structure.

Automated clustering for fragmented project notes groups related raw files based on content similarity and triage metadata. This intelligent grouping enables structured synthesis, proposing actions like merging, superseding, or copying content into a canonical format for final approval.

An initialized SweetClaude project environment is required to reconcile a corpus. You must also complete the corpus triage phase beforehand to ensure data integrity before the automated synthesis and subagent-driven drafting processes can begin.

Maintaining version control audit trails is fully supported when merging raw documentation. Every approved canonical document is tracked via Git commits, ensuring a clear and traceable history of project knowledge updates.

Subagent-driven drafting for knowledge management works by proposing how to merge, supersede, or copy clustered raw files into a canonical format. This structured synthesis automates the transition from raw research notes to approved project assets.

Automated synthesis should not be used before completing the corpus triage phase, as data integrity is required for accurate clustering. It is also not intended for unstructured repositories outside a defined SweetClaude project environment.