workflow-from-chats

To turn chat transcripts into reusable workflow guidance, provide recent chats and subagent transcripts. The system infers durable working preferences, clusters them by workflow shape, and generates artifacts like new skills, rules, or workflow documents.

Chat analysis extracts durable working preferences by scanning parent transcripts and relevant subagent transcripts. It uses parent conversations for citations, clusters findings by workflow shape, and excludes sensitive content to ensure safe artifact generation.

Yes, you can generate workflow rules from conversational feedback safely. The system explicitly excludes sensitive content during chat analysis and provides a concise evidence corpus with confidence ratings to support artifact adoption.

The best way to codify team practices from chat history is to infer durable preferences from recent chats, cluster them by workflow shape like shipping or review, and propose artifacts such as new skills, rules, or workflow documents.

To rate confidence when generating workflow artifacts from chats, the system evaluates inferred preferences against the scanned transcript evidence. It provides a concise evidence corpus with confidence ratings to support artifact adoption and future iterations.