What problem does it solve? Batch processing and metadata extraction miss the cross-cutting themes, participant trajectories, and unexpected parallels hidden across heterogeneous research corpora. This Skill performs holistic, roaming analysis of raw files to surface serendipitous connections that structured pipelines overlook. ## Core Features & Use Cases - Holistic Corpus Roaming: Reads raw files deeply, follows conceptual threads across groups, languages, source types, and participants. - Connection & Pattern Detection: Identifies concept evolution, cross-group parallels, unexpected contrasts, and linguistic bridges, flagging themes appearing in 3+ unexpected places. - Structured Serendipity Reports: Writes reports to agent_reports/ with connections found, patterns identified, proposed map updates, navigation logs, and Unicode sparkline discovery dashboards. - Use Case: After indexing interview transcripts and field notes, run this agent to discover that a theme appearing in one participant group unexpectedly recurs across three unrelated sources, then propose cross-references for the knowledge maps. ## Quick Start Ask the orchestrator to run the serendipity agent over the raw corpus to find hidden cross-source connections and write a serendipity report.
