fix: auto-create knowledge webhook + periodic pull fallback

hive-c0re now auto-creates the Forgejo push webhook for
internal/knowledge at startup (ensure_webhook). this is what was
missing — the webhook endpoint existed but was never registered in
forgejo, so merging iris's PR didn't trigger a pull.

also adds a periodic hourly pull as a fallback (and an immediate pull
at startup to reconcile commits that landed while c0re was offline).

fixes #1244.

hive-c0re/src/knowledge.rs

@@ -7,7 +7,9 @@
 //! to the bind-mounted path inside the container.
 //!
 //! hive-c0re maintains the local clone. A Forgejo webhook notifies it
-//! on push to main so agents always see an up-to-date snapshot.
+//! on push to main so agents always see an up-to-date snapshot. The
+//! webhook is auto-created by [`ensure_webhook`] at startup. A
+//! periodic pull in `main.rs` provides a fallback cadence.
 
 use anyhow::{Context, Result};
 
@@ -133,10 +135,77 @@ async fn seed_readme(core_token: &str) -> Result<()> {
     }
 }
 
+/// Ensure a Forgejo push webhook for `internal/knowledge` exists and
+/// points at hive-c0re's `/webhook/knowledge` endpoint. Idempotent —
+/// lists existing hooks first and skips creation when one is already
+/// targeting the correct URL. `dashboard_port` is the TCP port
+/// hive-c0re's dashboard listens on (default 7000); the webhook URL
+/// is `http://127.0.0.1:<port>/webhook/knowledge`.
+///
+/// Called at startup alongside [`ensure_local_clone`]. No-op when the
+/// core token is absent (forge not yet provisioned).
+pub async fn ensure_webhook(core_token: &str, dashboard_port: u16) -> Result<()> {
+    const FORGE_HTTP: &str = "http://localhost:3000";
+    let target_url = format!("http://127.0.0.1:{dashboard_port}/webhook/knowledge");
+    let client = reqwest::Client::builder()
+        .timeout(std::time::Duration::from_secs(10))
+        .build()
+        .context("build reqwest client for webhook setup")?;
+
+    // List existing hooks — skip creation if ours is already there.
+    let list_url = format!("{FORGE_HTTP}/api/v1/repos/{ORG}/{REPO}/hooks");
+    let resp = client
+        .get(&list_url)
+        .header("Authorization", format!("token {core_token}"))
+        .send()
+        .await
+        .with_context(|| format!("GET {list_url}"))?;
+    if resp.status().is_success() {
+        let hooks: Vec<serde_json::Value> = resp.json().await.unwrap_or_default();
+        let already_exists = hooks.iter().any(|h| {
+            h.get("config")
+                .and_then(|c| c.get("url"))
+                .and_then(|u| u.as_str())
+                == Some(&target_url)
+        });
+        if already_exists {
+            tracing::debug!(%target_url, "knowledge: push webhook already configured");
+            return Ok(());
+        }
+    }
+
+    // Create the webhook.
+    let create_url = format!("{FORGE_HTTP}/api/v1/repos/{ORG}/{REPO}/hooks");
+    let body = serde_json::json!({
+        "type": "forgejo",
+        "config": {
+            "url": target_url,
+            "content_type": "json"
+        },
+        "events": ["push"],
+        "active": true
+    });
+    let resp = client
+        .post(&create_url)
+        .header("Authorization", format!("token {core_token}"))
+        .json(&body)
+        .send()
+        .await
+        .with_context(|| format!("POST {create_url}"))?;
+    let status = resp.status();
+    if status.is_success() {
+        tracing::info!(%target_url, "knowledge: push webhook created");
+        Ok(())
+    } else {
+        let body = resp.text().await.unwrap_or_default();
+        anyhow::bail!("create webhook for {ORG}/{REPO} failed ({status}): {body}")
+    }
+}
+
 /// Pull the latest changes in the local clone. Called from the webhook
-/// handler on every push to `internal/knowledge` main. Uses
-/// `--ff-only` so a force-push to the knowledge repo never wedges the
-/// local copy silently.
+/// handler on every push to `internal/knowledge` main, and periodically
+/// from `main.rs` as a fallback. Uses `--ff-only` so a force-push to
+/// the knowledge repo never wedges the local copy silently.
 pub async fn pull() -> Result<()> {
     // Sanity: if the clone is missing (e.g. storage was wiped), refuse
     // to pull and let the caller decide whether to re-clone.

hive-c0re/src/main.rs

@@ -13,8 +13,8 @@ use hive_sh4re::{HostRequest, HostResponse};
 use hive_c0re::coordinator::Coordinator;
 use hive_c0re::{
     agent_sockets, auto_update, bash_tasks_vacuum, broker, client, crash_watch, dashboard,
-    dashboard_events, events_vacuum, forge, manager_server, matrix, migrate, rebuild_queue,
-    reminder_scheduler, scheduled_prompts_worker, server, stats_vacuum,
+    dashboard_events, events_vacuum, forge, knowledge, manager_server, matrix, migrate,
+    rebuild_queue, reminder_scheduler, scheduled_prompts_worker, server, stats_vacuum,
 };
 
 #[derive(Parser)]
@@ -262,6 +262,43 @@ async fn cmd_serve(
     tokio::spawn(async move {
         forge::ensure_all().await;
     });
+    // Knowledge webhook setup: ensure the Forgejo push webhook for
+    // `internal/knowledge` exists so `pull()` fires on merge. Runs
+    // after forge::ensure_all so the core token + repo are present.
+    // No-op when the core token or forge are absent.
+    let webhook_port = dashboard_port;
+    tokio::spawn(async move {
+        if let Some(token) = forge::core_token() {
+            if let Err(e) = knowledge::ensure_webhook(&token, webhook_port).await {
+                tracing::warn!(error = ?e, "knowledge: ensure_webhook failed");
+            }
+        }
+    });
+    // Knowledge periodic pull: hourly fallback in case the webhook is
+    // missed (e.g. hive-c0re was down during a push). First fires at
+    // startup (immediate pull after the clone is already present).
+    let mut knowledge_shutdown = coord.shutdown_rx();
+    tokio::spawn(async move {
+        // Initial pull — reconcile any commits that landed while c0re
+        // was offline, including the push that triggered #1244.
+        if let Err(e) = knowledge::pull().await {
+            tracing::debug!(error = ?e, "knowledge: startup pull skipped (no clone yet?)");
+        }
+        let interval = std::time::Duration::from_hours(1);
+        loop {
+            tokio::select! {
+                () = tokio::time::sleep(interval) => {
+                    if let Err(e) = knowledge::pull().await {
+                        tracing::warn!(error = ?e, "knowledge: periodic pull failed");
+                    }
+                }
+                _ = knowledge_shutdown.changed() => {
+                    tracing::info!("knowledge pull: shutdown signal received");
+                    break;
+                }
+            }
+        }
+    });
     // Matrix user sweep: same shape — ensure every container has
     // an account on the local matrix-tuwunel homeserver with an
     // access_token persisted to `<state>/matrix-token`. No-op when