wiki-importlisted

# wiki-import Bulk-import an existing document system into the wiki. Unlike `wiki-ingest` (which handles one source at a time), `wiki-import` scans an entire directory tree, infers structure, deduplicates against existing content, and processes files in waves with checkpoint support for large imports. ## When to use - User has an existing folder of notes, an Obsidian vault, a Notion export, a Confluence export, or any collection of markdown/text files - User says "import my existing docs", "migrate my notes", "bring in my whole vault" - Bootstrapping a wiki from an existing knowledge store ## Implementation Checkpoint persistence + import-lock IO — the parts that absolutely must not depend on agent self-discipline — live in `plugin/scripts/import_checkpoint.py`. **The script is the source of truth; the prose below explains its behavior.** The skill still does the LLM-heavy work (mapping inference, dedup judgment, page writing); the script owns durable state so a crash mid-import is recoverable AND wiki-sync can detect "import in progress" before reading a half-imported working tree. ```bash # Phase 1 — also acquire import lock (PRD-v1.8 §10/§11.8) python {plugin_root}/scripts/import_checkpoint.py --wiki {wiki_path} \ lock --source /path/to/notes --format obsidian python {plugin_root}/scripts/import_checkpoint.py --wiki {wiki_path} \ init --source /path/to/notes --format obsidian --total 342 # Update progress after each wave of 20 files python {plugin_root}/sc