docs: document crawlkit integration

CHANGELOG.md

@@ -2,6 +2,14 @@
 
 ## Unreleased
 
+- Depend on `github.com/vincentkoc/crawlkit v0.4.0` for shared config,
+  status/control, snapshot, mirror, output, and terminal explorer mechanics.
+- Keep Notion API/Desktop parsing, Markdown rendering, page/comment/database
+  schemas, Notion FTS body construction, and data-source compatibility
+  app-owned while the shared mechanics move to crawlkit.
+- Document the gitcrawl-style document TUI shape: workspace/teamspace/page or
+  database groups, page/database rows, preview/comment detail, sorting, mouse
+  selection, right-click actions, and local/remote status chrome.
 - Add crawlkit control metadata/status surfaces with `metadata --json`, `status --json`, and `doctor --json`.
 - Report primary archive and desktop-cache SQLite inventories in status JSON for shared local control surfaces.
 - Add `notcrawl tui`, a local terminal browser for archived pages and databases backed by `crawlkit/tui`.

README.md

@@ -77,6 +77,8 @@ Default paths:
 - `init` writes a starter config
 - `doctor` checks config, SQLite, desktop cache, and token presence
 - `status` prints archive counts, last sync time, and database/WAL size
+- `metadata --json`, `status --json`, and `doctor --json` expose crawlkit
+  control/status payloads for launchers, automation, and CI
 - `report` summarizes recent page, database, space, and comment activity
 - `maintain` rebuilds FTS, optimizes SQLite indexes, and can run `VACUUM`
 - `sync` ingests from `desktop`, `api`, or `all`
@@ -90,6 +92,20 @@ Default paths:
 - `subscribe` clones a share repo and imports the latest snapshot
 - `update` pulls and imports a subscribed share repo
 
+## Shared crawlkit surfaces
+
+`notcrawl` uses `crawlkit` for standard config paths, SQLite open/read helpers,
+snapshot packing/import, git-backed archive sharing, output formatting, status
+payloads, and the shared terminal explorer. Notion API/Desktop parsing,
+Markdown rendering, page/comment/database schemas, and Notion FTS bodies remain
+owned by `notcrawl`.
+
+The TUI follows the gitcrawl-style three-pane model: workspace/teamspace/page or
+database groups on the left, pages/databases in the middle, and a readable
+document preview plus comments and metadata on the right. It supports pane
+focus, sortable headers, mouse selection, right-click actions, and a
+local/remote footer.
+
 ## Distribution
 
 Release packaging is managed with GoReleaser. Tagged releases build tarballs,

docs/distribution.md

@@ -10,6 +10,15 @@ go test ./...
 go build ./cmd/notcrawl
 ```
 
+Also smoke the crawlkit control and non-interactive TUI surfaces before a tag:
+
+```bash
+notcrawl metadata --json
+notcrawl status --json
+notcrawl doctor --json
+notcrawl tui --json --limit 10
+```
+
 If GoReleaser is installed:
 
 ```bash