feat(tabs): reuse an open tab instead of duplicating a page

- Add browser_cli/url_match.py with a canonical URL key for tab matching
- Canonicalize scheme, host, default ports, www., trailing slash, query order
- Support reuse modes none, exact, similar and origin via normalize_mode()
- Pick the closest match when several tabs qualify, preferring identical URLs
- Skip the reload when the reused tab already shows the requested page
- Expose nav.reusable_tab() so callers can look up a tab without side effects
- Accept reuse/reuse_domain/reuse_title in tabs.open(), not just nav.open()
- Add --reuse-exact to nav open and nav open-wait; --reuse is now fuzzy
- Default MCP tabs_open to reuse="similar" and return a reused flag
- Tell MCP agents in the server instructions to reuse before opening
- Document the reuse modes in README for MCP, CLI and Python API
- Cover url_match, SDK reuse paths and MCP reuse behaviour with new tests
- Bump package and extension version to 0.16.9
- Refresh uv.lock for the updated dependency set

README.md

@@ -179,6 +179,21 @@ optional `tab_id` and act on the active tab when it is omitted, so a caller
 does not need a preceding `browser_tabs_list` round trip. `browser_tabs_close`
 reports the tab it closed.
 
+`browser_tabs_open` does not duplicate pages. It takes over an already-open tab
+whose URL is nearly the same and reports `reused` next to the tab metadata, so
+an agent that re-opens a page keeps working in the tab the user already has.
+The `reuse` argument selects the match:
+
+| `reuse` | Matches |
+|---------|---------|
+| `similar` (default) | Same site and page, ignoring fragment, trailing slash, query order, `www.` and default ports — plus any tab of the site when the requested URL is its root |
+| `exact` | Only a character-for-character identical URL |
+| `origin` | Any tab of the same site, whatever its path |
+| `none` | Never reuse, always open a new tab |
+
+A reused tab is only navigated when it is not already on the requested page, so
+no reload throws away page state.
+
 Available tools:
 - `browser_tabs_list`, `browser_tabs_open`, `browser_tabs_close`
 - `browser_navigate`, `browser_page_info`
@@ -244,6 +259,12 @@ browser-cli nav open https://example.com --focus            # bring opened tab/w
 browser-cli nav open https://example.com --window work      # into a named window
 browser-cli nav open https://example.com --group research   # into a tab group (name or ID)
 
+# Reuse an open tab instead of duplicating a page
+browser-cli nav open https://example.com/app --reuse         # nearly identical URL (fragment, trailing slash, query order, www)
+browser-cli nav open https://example.com/app --reuse-exact   # only a character-for-character match
+browser-cli nav open https://example.com/app --reuse-domain  # any tab of the same site
+browser-cli nav open https://example.com/app --reuse-title "App"  # tab whose title contains the text
+
 # Reload
 browser-cli nav reload                  # reload active tab
 browser-cli nav reload 1234             # reload tab by ID
@@ -456,6 +477,9 @@ b.nav.search("google", "python asyncio")
 tabs = b.tabs.list()          # list[Tab]; in multi-browser mode each tab.browser is set
 tab = b.tabs.open("https://example.com")              # returns a bound Tab object
 tab = b.tabs.open("https://example.com", wait=True, timeout=10)
+tab = b.tabs.open("https://example.com/app", reuse=True)   # reuse a tab with a nearly identical URL
+tab = b.tabs.open("https://example.com/app", reuse="origin")  # reuse any tab of that site
+existing = b.nav.reusable_tab("https://example.com/app", reuse=True)  # Tab or None, without touching the browser
 active = b.tabs.active()      # active Tab object
 tab = b.tabs.get(1234)        # tab by ID
 tab = b.tabs.first("github")  # first matching tab or None