diff --git a/docs/tools/forge.md b/docs/tools/forge.md
index f3dead35..53b5cb51 100644
--- a/docs/tools/forge.md
+++ b/docs/tools/forge.md
@@ -37,7 +37,7 @@ hive-forge pr-reviews 42 --comment -m "msg"  # submit COMMENT review
 hive-forge diff 42                          # unified diff (lockfile hunks collapsed by default)
 hive-forge diff 42 --full                   # include unfiltered lockfile hunks
 hive-forge list                             # open issues/PRs
-hive-forge list --kind pr --state all --all # every PR across all pages (repo-wide sweep; --all follows pagination)
+hive-forge list --kind pr --state all --page 2  # page 2 of all PRs (walk --page 1,2,… with --limit as page size for a repo-wide sweep)
 hive-forge milestone                        # list milestones
 hive-forge branches deployed/               # filter branches by pattern
 hive-forge tree-sha main                    # git tree SHA for a ref
diff --git a/hive-forge/src/verbs/list.rs b/hive-forge/src/verbs/list.rs
index 9cac82df..01197652 100644
--- a/hive-forge/src/verbs/list.rs
+++ b/hive-forge/src/verbs/list.rs
@@ -1,8 +1,9 @@
 //! `list [--kind issue|pr|both] [--state open|closed|all] [--assignee
-//! <user>] [--author <user>] [--label <name>] [--limit N] [--all]` —
+//! <user>] [--author <user>] [--label <name>] [--limit N] [--page N]` —
 //! list issues / PRs with filters. Pretty `#NNN [author] title` output by
-//! default; `--json` for piping. `--all` drains every page (for repo-wide
-//! analysis) instead of a single `--limit` page.
+//! default; `--json` for piping. `--limit` is the page size and `--page`
+//! the 1-based page number — walk pages incrementally for a large result
+//! set rather than pulling the whole population into one response.
 //!
 //! Mirrors Forgejo's `GET /repos/{owner}/{repo}/issues` query-string
 //! filters one-for-one so the mental model carries over. Closes the
@@ -84,29 +85,27 @@ pub struct Args {
     /// Filter to items carrying any of these label names. Repeatable.
     #[arg(long = "label")]
     labels: Vec<String>,
-    /// Max items to return (default: 30; forge's per-page cap applies).
+    /// Page size — items per page (default: 30; forge's per-page cap,
+    /// ~50, applies).
     #[arg(long, default_value_t = 30)]
     limit: u64,
-    /// Drain ALL matching items across every page, not just the first
-    /// `--limit` (which the forge caps at its per-page max, ~50). Follows
-    /// pagination up to a safety cap. Use for repo-wide analysis — e.g.
-    /// `--kind pr --state all --all` to sweep every PR. Overrides
-    /// `--limit` (each page is fetched at the per-page max).
-    #[arg(long)]
-    all: bool,
+    /// Page number to fetch (1-based; default 1). Combine with `--limit`
+    /// to walk a large result set incrementally — fetch page 1, process,
+    /// fetch page 2, … until a short/empty page. This keeps each call
+    /// token-bounded (one page at a time) instead of pulling a whole
+    /// repo's population into a single response.
+    #[arg(long, default_value_t = 1)]
+    page: u64,
 }
 
 pub fn run(client: &Client, args: Args) -> Result<()> {
     let repo = client.repo();
-    // With `--all` each page is fetched at the forge's per-page max so we
-    // make the fewest round-trips; `get_json_all` then follows the
-    // `Link: rel="next"` chain. Otherwise honour `--limit` as a single page.
-    let per_page = if args.all { 50 } else { args.limit };
     let mut path = format!(
-        "/repos/{repo}/issues?type={}&state={}&limit={}",
+        "/repos/{repo}/issues?type={}&state={}&limit={}&page={}",
         args.kind.api_value(),
         args.state.api_value(),
-        per_page
+        args.limit,
+        args.page
     );
     if let Some(u) = args.assignee.as_deref()
         && !u.is_empty()
@@ -130,20 +129,6 @@ pub fn run(client: &Client, args: Args) -> Result<()> {
         let encoded: Vec<String> = args.labels.iter().map(|l| super::pct_encode(l)).collect();
         write!(path, "&labels={}", encoded.join(",")).unwrap();
     }
-    if args.all {
-        // Safety cap: 200 pages * 50/page = 10k items, well past any
-        // realistic repo while still bounding a runaway loop.
-        const MAX_PAGES: u32 = 200;
-        let items = client.get_json_all(&path, MAX_PAGES)?;
-        if client.json_mode() {
-            return print_json(&Value::Array(items));
-        }
-        for item in &items {
-            print_row(item);
-        }
-        return Ok(());
-    }
-
     let resp = client.get_json(&path)?;
     if client.json_mode() {
         return print_json(&resp);