workflow-from-chats

To extract workflow preferences from chat transcripts, the skill inventories recent conversations, identifies explicit rules and workflow markers, and synthesizes them into actionable preference atoms with confidence scoring. It builds a preference profile from prior work without manual sorting.

Yes, you can convert chat history into reusable CI review guidance by mining recent transcripts for consistent triggers, quality bars, and stop conditions. The skill clusters preferences by workflow shape to generate standardized review rules for agents.

Standardizing agent workflows from prior conversations involves extracting durable preferences from chat transcripts and selecting the appropriate artifact format like a skill, rule, or workflow doc. The skill uses evidence-driven synthesis to match established team patterns.

Preference extraction handles contradicting workflow evidence by assigning weak or contradicted confidence levels to the conflicting preference atoms. When evidence is blocked or contradictory, the skill defers artifact creation and outputs open questions.

You do not need specific chat formats to mine workflow preferences; the skill processes recent chat transcripts directly from your environment. It applies transcript inventorying and preference atom extraction to standard text conversation logs without structured data inputs.