Workflow From Chats

This one mines your recent Cursor chats to extract repeatable workflow preferences and turn them into skills, rules, or docs. It doesn't summarize conversations. It looks for patterns like corrections, explicit preferences ("I always want X"), and recurring workflows around shipping, debugging, or reviews. The confidence scoring is honest: strong for direct requests or repeated corrections, weak for one-off behavior, contradicted when evidence conflicts. Useful when you realize you keep giving the same feedback or want to codify how your team actually works. Defaults to the last week of transcripts and won't leak paths or secrets into artifacts.

skills.sh installs: 2.2K
GitHub stars: 15.9K
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add cursor/plugins --skill workflow-from-chats --agent claude-code

Installs into .claude/skills of the current project.

Files

Workflow From Chats

Infer durable working preferences from recent chats. Do not summarize chats; extract reusable workflow guidance.

Scope

Default to the last 7 days unless the user asks for a different window.
Read parent transcripts and relevant subagent transcripts. Use subagent content as evidence, but cite only parent conversations.
Do not expose local transcript paths, secrets, customer data, private chat content, or credentials.