conversation-knowledge-flywheel

What problem does it solve?

Many AI conversation transcripts remain ephemeral and unlocked; this skill turns raw chat transcripts into discoverable, project-anchored knowledge notes, prioritized next-day suggestions, and writer-ready packages so insights become actionable and visible in the Brain.

Core Features & Use Cases

• Manifest generation that inventories transcript files and infers project groupings.
• High-recall candidate retrieval via a local QMD collection layer with a lightweight Surveillance scan to surface priority candidates.
• Project routing to the Brain-side 05-PROJECTS registry, writer-package rendering, and commit/visibility verification so notes become Obsidian-visible.
• Failure and degraded-mode handling: QMD healthchecks, repair attempts, and explicit degraded-run reports.
• Use case: run nightly to convert yesterday's AI conversations into 1-3 knowledge drafts, a daily suggestions block, and optional research seed candidates for NotebookLM reinforcement.

Quick Start

Run the conversation-knowledge-flywheel for the previous day to generate the transcript manifest, surveillance shortlist, and writer-ready drafts for the Brain.