What problem does it solve? Pasted ChatGPT conversations are ephemeral and hard to reuse: decisions, reasoning, reusable prompts, and file references get lost or flattened when copied out of the chat UI. This Skill turns manually supplied ChatGPT material into a durable, evidence-aware Markdown extract that remains understandable and actionable without access to the original thread, account, or Project. ## Core Features & Use Cases - Evidence-aware extraction: Builds turn ledgers and content element ledgers, separates stated facts from inferences and unknowns, and records capture mode, completeness, and provenance for every artifact. - Depth-controlled distillation: Supports rapid, balanced, and comprehensive extraction profiles (with aliases like scan, distill, and catalog) plus focus, must-preserve, and safe-exclusion controls. - Deterministic artifact writer: A bundled Python script validates metadata, derives a filesystem-safe slug from a three-pass title chain, prevents accidental overwrites, and writes the final Markdown file atomically. - Use Case: You paste a long ChatGPT Project conversation including a Canvas document and citations. The Skill normalizes speaker turns, catalogs the Canvas and citations as elements, extracts decisions and next actions, and writes a standalone handoff document to docs/thread-extracts/. ## Quick Start Paste your ChatGPT conversation and ask the agent to extract it into a standalone Markdown artifact at balanced depth using this Skill.
