session-to-templates

To extract reusable prompt templates from Claude Code sessions, this Skill analyzes stored session files, scores exchanges, samples key turns, and uses an LLM to distill structured templates into a reusable library.

Building a prompt library from LLM session history is done by mining high-signal turns using head, tail, and key-turn sampling to preserve problem-solution context, then generating structured prompts stored in a manifest.

Yes, session-to-templates supports resume-capable processing for large session files by storing extracted outputs with metadata in a manifest, enabling deduplication and allowing batch processing to resume.

Yes, you need an LLM API key configured via python-dotenv to generate templates from AI session logs, as the Skill leverages an LLM to analyze sampled turns and distill them into structured reusable prompts.

Yes, you can automate prompt extraction for common coding tasks in Claude Code by running the batch tool, which scans session files, scores exchanges, and generates a structured template library to speed up new projects.

Key-turn sampling when mining prompts from conversation logs works by selecting head, tail, and critical middle exchanges to preserve problem-solution context while keeping the distilled LLM prompt templates concise and reusable.