fold forge_cursor into hyperhive-harness.json + prose comments (mara/argus review)

docs/forge.md

@@ -92,17 +92,32 @@ that unread signal would be consumed before the agent acts and the
 guard could never fire.
 
 Because a delivered thread stays unread, it reappears in every
-`?all=false` poll. An in-memory **delivery-dedupe cursor** (thread
+`?all=false` poll. A **delivery-dedupe cursor** (thread id →
-id → last-delivered `updated_at`, held in the poll loop) stops the
+last-delivered `updated_at`) stops the same version from re-firing a
-same version from re-firing a wake; a new comment bumps `updated_at`
+wake; a new comment bumps `updated_at` so genuinely new activity
-so genuinely new activity re-delivers. The cursor is pure anti-spam,
+re-delivers. The cursor is pure anti-spam, not a correctness oracle.
-not a correctness oracle: lost on harness restart it just
+Each poll prunes it to the threads still in the unread set. A failed
-re-delivers currently-unread threads once (harmless — `recv`
+wake delivery is left unread **and** out of the cursor, so it
-tolerates redelivery), so it carries none of the persisted-mirror
+resurfaces next tick.
-fragility that ruled out an on-disk seen-cursor. Each poll prunes
+
-the cursor to the threads still in the unread set. A failed wake
+The cursor is **persisted** as the `forge_cursor` field of the
-delivery is left unread **and** out of the cursor, so it resurfaces
+harness's consolidated `hyperhive-harness.json` state file (atomic
-next tick.
+tmp+rename, flushed only when it changed) and reloaded on boot, so a
+container rebuild/restart doesn't re-deliver the whole currently-unread
+backlog (#2106 — previously the in-memory-only cursor was lost on
+restart and every old still-unread thread re-fired a wake). The poller
+runs in the same harness process that owns that file, so it's one
+daemon → one state file rather than a second json; both writers
+(turn-loop fields + this cursor) go read-modify-write under a shared
+lock so neither clobbers the other's fields. This is safe because a
+thread is recorded **after** a successful broker delivery, and the
+broker inbox is durable sqlite — so a persisted "delivered" entry can
+never swallow a wake the agent never received. Crucially the cursor is
+a private dedup mirror, **not** forge's read-state: it does not
+reintroduce the read-before-comment coupling that ruled out the old
+mark-read-on-delivery approach. A missing (first boot) or malformed
+cursor degrades to empty — re-deliver the unread set once — never an
+abort.
 
 Self-echo notifications (the agent's own writes, see below) are the
 one path still marked-read directly (no read-before-comment value).

docs/persistence.md

@@ -129,7 +129,7 @@ Consolidated harness state file written atomically (`.tmp` + rename) by
 Shape:
 
 ```json
-{ "rate_limited": false, "needs_login": false }
+{ "rate_limited": false, "needs_login": false, "active_model": "…", "forge_cursor": { "42": "2026-07-01T18:00:00Z" } }
 ```
 
 - `rate_limited` — set when the harness detects a 429 from the Claude
@@ -138,6 +138,17 @@ Shape:
 - `needs_login` — set when a turn hits 401 (expired OAuth credentials);
   cleared by `"online"` status (re-auth completed). Drives the
   `needs_login` flag alongside the `claude_has_session` check.
+- `active_model` — the resolved Claude model for the dashboard badge.
+- `forge_cursor` — the `forge_notify` delivery-dedupe cursor
+  (notification thread id → last-delivered `updated_at`), so a
+  rebuild/restart doesn't re-deliver the whole currently-unread forge
+  backlog. See [`forge.md`](forge.md).
+
+Multiple harness tasks write this file (the turn loop for the first
+three fields, the `forge_notify` poller for `forge_cursor`), so every
+writer goes read-modify-write under a shared in-process lock — each
+preserves the fields it doesn't own rather than reconstructing the file
+from scratch.
 
 hive-c0re reads this file on each `build_all` sweep (~10s) via
 `container_view::read_harness_flags`. Falls back to the legacy individual

hive-ag3nt/src/events.rs

@@ -106,6 +106,34 @@ fn harness_json_path() -> PathBuf {
     crate::paths::state_dir().join(HARNESS_JSON)
 }
 
+// Serialises the read-modify-write of `hyperhive-harness.json`. Two
+// harness tasks touch it in the same process — the turn loop (rate-limit
+// / needs-login / active-model) and the forge_notify poller (the
+// delivery-dedupe cursor) — writing disjoint fields, so each writer must
+// preserve the other's. The lock closes the lost-update window between a
+// writer's read and its rename.
+static HARNESS_JSON_LOCK: std::sync::Mutex<()> = std::sync::Mutex::new(());
+
+/// Read the consolidated state file as a JSON object, or an empty object
+/// when it is missing / unparseable / not an object.
+fn read_harness_json() -> serde_json::Value {
+    std::fs::read_to_string(harness_json_path())
+        .ok()
+        .and_then(|raw| serde_json::from_str::<serde_json::Value>(&raw).ok())
+        .filter(serde_json::Value::is_object)
+        .unwrap_or_else(|| serde_json::json!({}))
+}
+
+/// Atomically overwrite the state file (`.tmp` + rename) so hive-c0re
+/// never reads a partial file.
+fn write_harness_json(v: &serde_json::Value) {
+    let path = harness_json_path();
+    let tmp = path.with_extension("json.tmp");
+    if std::fs::write(&tmp, v.to_string()).is_ok() {
+        let _ = std::fs::rename(&tmp, &path);
+    }
+}
+
 fn read_harness_state() -> (bool, bool, Option<String>) {
     // Try the new consolidated file first.
     if let Ok(raw) = std::fs::read_to_string(harness_json_path())
@@ -133,24 +161,55 @@ fn read_harness_state() -> (bool, bool, Option<String>) {
     (rate_limited, needs_login, None)
 }
 
-/// Write harness state atomically via a `.tmp` + `rename` pair so
+/// Write the turn-loop's harness state fields via a read-modify-write so
-/// hive-c0re never reads a partial file. Pass `active_model: Some(s)`
+/// any other writer's fields (e.g. `forge_notify`'s `forge_cursor`) survive.
-/// to include the resolved model (as surfaced in the dashboard badge);
+/// Pass `active_model: Some(s)` to update the resolved model (surfaced in
-/// `None` omits the field, which hive-c0re treats as "not yet known".
+/// the dashboard badge); `None` leaves the stored value untouched.
 fn write_harness_state(rate_limited: bool, needs_login: bool, active_model: Option<&str>) {
-    let path = harness_json_path();
+    let _guard = HARNESS_JSON_LOCK
-    let mut json = serde_json::json!({
+        .lock()
-        "rate_limited": rate_limited,
+        .unwrap_or_else(std::sync::PoisonError::into_inner);
-        "needs_login": needs_login,
+    let mut v = read_harness_json();
-    });
+    v["rate_limited"] = rate_limited.into();
+    v["needs_login"] = needs_login.into();
     if let Some(model) = active_model {
-        json["active_model"] = serde_json::Value::String(model.to_string());
+        v["active_model"] = model.into();
-    }
-    let body = json.to_string();
-    let tmp = path.with_extension("json.tmp");
-    if std::fs::write(&tmp, &body).is_ok() {
-        let _ = std::fs::rename(&tmp, &path);
     }
+    write_harness_json(&v);
+}
+
+/// Parse the `forge_notify` delivery-dedupe cursor (notification thread id
+/// -> last-delivered `updated_at`) out of a harness-state JSON value.
+/// Empty when the field is absent (first boot) or malformed.
+fn forge_cursor_from_json(v: &serde_json::Value) -> std::collections::HashMap<u64, String> {
+    v.get("forge_cursor")
+        .cloned()
+        .and_then(|c| serde_json::from_value(c).ok())
+        .unwrap_or_default()
+}
+
+/// Restore the `forge_notify` delivery-dedupe cursor from the consolidated
+/// state file so a container rebuild/restart doesn't re-deliver the whole
+/// currently-unread backlog.
+pub fn read_forge_cursor() -> std::collections::HashMap<u64, String> {
+    forge_cursor_from_json(&read_harness_json())
+}
+
+/// Persist the `forge_notify` delivery-dedupe cursor into the consolidated
+/// state file, read-modify-write under the shared lock so the turn-loop's
+/// own fields survive. Best-effort: a serialize failure is a no-op.
+pub fn write_forge_cursor<S: std::hash::BuildHasher>(
+    cursor: &std::collections::HashMap<u64, String, S>,
+) {
+    let Ok(value) = serde_json::to_value(cursor) else {
+        return;
+    };
+    let _guard = HARNESS_JSON_LOCK
+        .lock()
+        .unwrap_or_else(std::sync::PoisonError::into_inner);
+    let mut v = read_harness_json();
+    v["forge_cursor"] = value;
+    write_harness_json(&v);
 }
 
 fn now_unix() -> i64 {
@@ -1145,10 +1204,37 @@ impl Default for Bus {
 mod tests {
     use super::{
         BusEvent, DEFAULT_EFFORT, EFFORT_LEVELS, LiveEvent, StoredEvent, TokenUsage,
-        is_valid_effort,
+        forge_cursor_from_json, is_valid_effort,
     };
     use serde_json::json;
 
+    #[test]
+    fn forge_cursor_absent_field_is_empty() {
+        // First boot / a state file that only carries the turn-loop fields:
+        // no cursor yet, so we re-deliver the currently-unread set once.
+        assert!(forge_cursor_from_json(&json!({ "rate_limited": false })).is_empty());
+    }
+
+    #[test]
+    fn forge_cursor_malformed_field_is_empty() {
+        // A wrong-typed / corrupt cursor degrades to empty rather than
+        // aborting the poller.
+        assert!(forge_cursor_from_json(&json!({ "forge_cursor": "nonsense" })).is_empty());
+    }
+
+    #[test]
+    fn forge_cursor_roundtrips_u64_keys() {
+        // serde_json stringifies integer map keys; confirm the u64 thread
+        // ids the cursor is keyed on survive the JSON round-trip.
+        let v = json!({
+            "forge_cursor": { "42": "2026-06-22T16:00:00Z", "99": "2026-06-22T17:30:00Z" }
+        });
+        let cursor = forge_cursor_from_json(&v);
+        assert_eq!(cursor.len(), 2);
+        assert_eq!(cursor.get(&42), Some(&"2026-06-22T16:00:00Z".to_owned()));
+        assert_eq!(cursor.get(&99), Some(&"2026-06-22T17:30:00Z".to_owned()));
+    }
+
     #[test]
     fn stored_event_serializes_ts_beside_kind() {
         // History-row wire shape: `ts` is a flattened sibling of `kind`,

hive-ag3nt/src/forge_notify.rs

@@ -4,10 +4,15 @@
 //! delivers it to the agent's inbox. Delivered threads are deliberately
 //! left UNREAD in forge — the hive-forge read-before-comment guard keys
 //! off forge's own unread-state, and the agent reading the thread via the
-//! CLI is what marks it read. An in-memory delivery-dedupe cursor
+//! CLI is what marks it read. A delivery-dedupe cursor (thread id →
-//! (thread id → last-delivered `updated_at`) stops the still-unread
+//! last-delivered `updated_at`) stops the still-unread notification from
-//! notification from re-firing a wake every poll; self-echo and
+//! re-firing a wake every poll; self-echo and drop-listed notifications
-//! drop-listed notifications are still marked read directly.
+//! are still marked read directly. The cursor is persisted as the
+//! `forge_cursor` field of the harness's consolidated `hyperhive-harness.json`
+//! (via [`crate::events`]) and reloaded on boot so a container
+//! rebuild/restart doesn't re-deliver the whole currently-unread backlog —
+//! it is a private dedup mirror, NOT forge's read-state, so the
+//! read-before-comment guard is untouched.
 //!
 //! Activation gates, self-notification filtering, body excerpt +
 //! truncation + heading escape, wrapper formats (comment / review /
@@ -123,15 +128,28 @@ pub async fn run(socket: PathBuf) {
     // Delivery-dedupe cursor: notification thread id -> the `updated_at`
     // of the version we last woke the agent for. We no longer mark a
     // thread read on delivery (that would consume the unread signal the
-    // hive-forge read-before-comment guard relies on), so this in-memory
+    // hive-forge read-before-comment guard relies on), so this map is what
-    // map is what stops the same unread notification from re-firing a
+    // stops the same unread notification from re-firing a wake every poll.
-    // wake every poll. A new comment bumps `updated_at`, so the thread
+    // A new comment bumps `updated_at`, so the thread re-delivers. This is
-    // re-delivers. This is purely anti-spam, NOT a correctness oracle:
+    // purely anti-spam, NOT a correctness oracle.
-    // lost on harness restart it just re-delivers currently-unread
+    //
-    // threads once (harmless — recv tolerates redelivery), so it carries
+    // It is persisted as the `forge_cursor` field of the harness's
-    // none of the persisted-mirror fragility that sank the on-disk
+    // consolidated state file and reloaded here on boot so a container
-    // cursor approach.
+    // rebuild/restart doesn't re-deliver the entire currently-unread
-    let mut delivered: HashMap<u64, String> = HashMap::new();
+    // backlog. Persisting is safe because we only record a thread AFTER a
+    // successful broker delivery, and the broker inbox is durable sqlite —
+    // so a persisted "delivered" entry can never swallow a wake the agent
+    // never received. The cursor is a private dedup mirror, decoupled from
+    // forge's own read-state, so it doesn't reintroduce the
+    // read-before-comment coupling that the mark-read-on-delivery approach
+    // suffered.
+    let mut delivered: HashMap<u64, String> = crate::events::read_forge_cursor();
+    if !delivered.is_empty() {
+        info!(
+            entries = delivered.len(),
+            "forge_notify: restored delivery-dedupe cursor"
+        );
+    }
 
     loop {
         interval.tick().await;
@@ -761,6 +779,11 @@ async fn poll_once(
         "forge_notify: delivering notifications"
     );
 
+    // Tracks whether the dedupe cursor changed this poll (a new delivery
+    // recorded, or the prune below dropped now-read threads) so we only
+    // rewrite the on-disk cursor when there's something to persist.
+    let mut cursor_dirty = false;
+
     for notif in &notifications {
         let Some(id) = notif["id"].as_u64() else {
             continue;
@@ -804,6 +827,7 @@ async fn poll_once(
                 // here in the Ok arm — a failed delivery leaves the cursor
                 // untouched, so it re-delivers next tick.
                 delivered.insert(id, updated_at);
+                cursor_dirty = true;
             }
             Err(e) => {
                 warn!(%id, error = ?e, "forge_notify: deliver failed — leaving unread");
@@ -821,7 +845,18 @@ async fn poll_once(
         .iter()
         .filter_map(|n| n["id"].as_u64())
         .collect();
+    let before_prune = delivered.len();
     delivered.retain(|id, _| current_ids.contains(id));
+    if delivered.len() != before_prune {
+        cursor_dirty = true;
+    }
+
+    // Flush the cursor to the consolidated state file only when it
+    // changed, so a rebuild/restart reloads it instead of re-delivering
+    // the whole unread backlog.
+    if cursor_dirty {
+        crate::events::write_forge_cursor(delivered);
+    }
 }
 
 /// Whether a notification should be delivered as a wake given the