What problem does it solve? Your OpenClaw agent accumulates valuable knowledge across sessions—decisions, patterns, project context—but it stays locked in MEMORY.md files, daily notes, and JSONL transcripts. This Skill mines that history and distills it into durable, organized Obsidian wiki pages instead of letting insights fade into append-only logs. ## Core Features & Use Cases - Multi-Source Ingestion: Parses MEMORY.md, daily notes (workspace/memory/YYYY-MM-DD.md), DREAMS.md, and per-session JSONL transcripts, ranked by signal value. - Incremental Sync: Uses a .manifest.json delta check to process only new or modified files in append mode, with a full re-ingest mode for rebuilds. - Privacy-First Distillation: Redacts credentials and sensitive data, summarizes rather than quoting transcripts, and clusters knowledge by topic rather than dumping one page per session. - Use Case: After months of OpenClaw sessions across multiple projects, run the ingest to surface recurring debugging patterns into skills/ pages, project decisions into projects/<name>/ pages, and cross-session insights into synthesis/ pages—complete with provenance markers and confidence metadata. ## Quick Start Ask the assistant to process your OpenClaw history and add the extracted knowledge to your Obsidian wiki.
