@@ -82,13 +82,19 @@ pub fn format_linked_issue(issue: &LinkedIssue) -> String {
     }
 }
 
-/// Format the Detail-generated fix PR (number, state, and URL) for the
-/// detail/show view, rendered as `#<number> (<state>) — <url>`.
+/// Format the Detail-generated fix PR for the detail/show view.
+///
+/// Rendered as `#<number> (<state>) — <url>`. The URL and its separator are
+/// omitted when the API returns an empty URL.
 pub fn format_fix_pr(fix_pr: &FixPr) -> String {
-    format!(
-        "#{} ({}) \u{2014} {}",
-        fix_pr.pr_number, fix_pr.state, fix_pr.url
-    )
+    if fix_pr.url.is_empty() {
+        format!("#{} ({})", fix_pr.pr_number, fix_pr.state)
+    } else {
+        format!(
+            "#{} ({}) \u{2014} {}",
+            fix_pr.pr_number, fix_pr.state, fix_pr.url
+        )
+    }
 }
 
 // ── clap::ValueEnum ──────────────────────────────────────────────────
@@ -502,6 +508,17 @@ mod tests {
         );
     }
 
+    #[test]
+    fn format_fix_pr_omits_separator_when_url_empty() {
+        let fix_pr: FixPr = serde_json::from_value(serde_json::json!({
+            "prNumber": 42,
+            "url": "",
+            "state": "merged"
+        }))
+        .expect("valid FixPr JSON");
+        assert_eq!(format_fix_pr(&fix_pr), "#42 (merged)");
+    }
+
     // ── format_introduced_in ─────────────────────────────────────────
 
     #[test]
 
@@ -11,7 +11,7 @@ use crate::api::types::{
     review_state_label, Bug, BugDismissalReason, BugId, BugReviewState,
     ListPublicBugsWorkflowRequestId, RepoId,
 };
-use crate::output::{output_list, SectionRenderer};
+use crate::output::{clamp_page, output_list, SectionRenderer};
 use crate::utils::datetime::{format_datetime, parse_time_spec};
 use crate::utils::pagination::page_to_offset;
 use crate::utils::repos::resolve_repo_id;
@@ -574,8 +574,9 @@ pub async fn handle(command: &BugCommands, cli: &crate::Cli) -> Result<()> {
                     let effective_limit = u32::try_from(total.max(1)).unwrap_or(u32::MAX);
                     return output_list(&filtered, total, 1, effective_limit, format);
                 }
-                let page_items = paginate_items(&filtered, *page, *limit);
-                output_list(&page_items, total, *page, *limit, format)
+                let page = clamp_page(*page, total, *limit);
+                let page_items = paginate_items(&filtered, page, *limit);
+                output_list(&page_items, total, page, *limit, format)
             } else if multi_status {
                 // Multiple statuses but no client-side filters: fetch one
                 // page per status and merge, avoiding a full exhaust.
 
@@ -3,7 +3,7 @@ use clap::Subcommand;
 
 use crate::api::client::ApiClient;
 use crate::api::types::{RepoId, Scan, ScanType, ScansResponse, WorkflowStatus};
-use crate::output::output_list;
+use crate::output::{clamp_page, output_list};
 use crate::utils::datetime::parse_time_spec;
 use crate::utils::pagination::page_to_offset;
 use crate::utils::repos::resolve_repo_id;
@@ -157,8 +157,9 @@ pub async fn handle(command: &ScanCommands, cli: &crate::Cli) -> Result<()> {
                     until_ms,
                 );
                 let total = filtered.len();
-                let page_items = paginate_items(&filtered, *page, *limit);
-                output_list(&page_items, total, *page, *limit, format)
+                let page = clamp_page(*page, total, *limit);
+                let page_items = paginate_items(&filtered, page, *limit);
+                output_list(&page_items, total, page, *limit, format)
             } else {
                 let offset = page_to_offset(*page, *limit);
                 let scans = client
 
@@ -118,6 +118,14 @@ fn total_pages(total: usize, limit: u32) -> u32 {
         .max(1)
 }
 
+/// Clamp a page number to the last page that exists.
+///
+/// Client-side filtering can shrink a result set below the requested page,
+/// which would otherwise render an impossible `Page: 2 of 1`.
+pub fn clamp_page(page: u32, total: usize, limit: u32) -> u32 {
+    page.clamp(1, total_pages(total, limit))
+}
+
 /// Generic helper to output a list of items in the requested format
 pub fn output_list<T: Formattable + Serialize>(
     items: &[T],
@@ -197,6 +205,28 @@ mod tests {
         assert_eq!(total_pages(10, 0), 1);
     }
 
+    // ── clamp_page ───────────────────────────────────────────────────
+
+    #[test]
+    fn clamp_page_keeps_valid_page() {
+        assert_eq!(clamp_page(2, 100, 50), 2);
+    }
+
+    #[test]
+    fn clamp_page_lowers_page_beyond_last() {
+        assert_eq!(clamp_page(5, 10, 50), 1);
+    }
+
+    #[test]
+    fn clamp_page_empty_results_is_first_page() {
+        assert_eq!(clamp_page(3, 0, 50), 1);
+    }
+
+    #[test]
+    fn clamp_page_raises_zero_to_first_page() {
+        assert_eq!(clamp_page(0, 100, 50), 1);
+    }
+
     // ── SectionRenderer builder ──────────────────────────────────────
 
     #[test]
 
@@ -26,11 +26,18 @@ pub fn repo_root() -> Result<PathBuf> {
 }
 
 /// Extract the `origin` remote's URL from `jj git remote list` output,
-/// which prints one `name url` pair per line.
+/// which prints one `name url` pair per line. When a remote's fetch and push
+/// URLs differ, jj appends a ` (push: <url>)` annotation that is dropped here
+/// so only the fetch URL is returned.
 fn parse_jj_remote_list(list: &str) -> Option<String> {
     list.lines()
         .find_map(|line| line.strip_prefix("origin "))
-        .map(|url| url.trim().to_string())
+        .map(|url| {
+            url.split_once(" (push:")
+                .map_or(url, |(fetch, _)| fetch)
+                .trim()
+                .to_string()
+        })
 }
 
 /// Extract `owner/repo` from a GitHub remote URL.
@@ -94,7 +101,9 @@ fn strip_http_credentials(url: &str) -> Option<String> {
             // (or end of string). Anything after the authority — including a
             // `@` in the path — stays untouched.
             let (authority, path) = rest.split_once('/').unwrap_or((rest, ""));
-            if let Some((_, host)) = authority.split_once('@') {
+            // The credentials/host separator is the *last* `@`, since a
+            // password or token may itself contain `@`.
+            if let Some((_, host)) = authority.rsplit_once('@') {
                 return Some(format!("{scheme}{host}/{path}"));
             }
         }
@@ -205,6 +214,17 @@ mod tests {
         );
     }
 
+    #[test]
+    fn jj_remote_list_strips_push_url_annotation() {
+        // jj appends `(push: <url>)` when fetch and push URLs differ.
+        assert_eq!(
+            parse_jj_remote_list(
+                "origin https://github.com/usedetail/cli.git (push: https://other.com/cli.git)"
+            ),
+            Some("https://github.com/usedetail/cli.git".to_string()),
+        );
+    }
+
     // ── parse_github_remote_url ─────────────────────────────────────
 
     #[test]
@@ -330,6 +350,15 @@ mod tests {
         );
     }
 
+    #[test]
+    fn parses_https_with_at_sign_in_password() {
+        // The credentials/host separator is the last `@`, not the first.
+        assert_eq!(
+            parse_github_remote_url("https://user:p@ss@github.com/usedetail/cli.git"),
+            Some("usedetail/cli".to_string()),
+        );
+    }
+
     #[test]
     fn parses_http_with_credentials() {
         assert_eq!(