@ouro.bot/cli 0.1.0-alpha.728 → 0.1.0-alpha.729

package/changelog.json

@@ -1,6 +1,13 @@
 {
   "_note": "This changelog is maintained as part of the PR/version-bump workflow. Agent-curated, not auto-generated. Agents read this file directly via read_file to understand what changed between versions.",
   "versions": [
+    {
+      "version": "0.1.0-alpha.729",
+      "changes": [
+        "Respawn long-lived auto-start workers after unsolicited clean exits so managed BlueBubbles listeners recover without a full daemon restart.",
+        "Bound repeated clean-exit loops with cause-neutral diagnostics and keep explicit stop, whole-daemon stop, and manual restart free of stale respawns."
+      ]
+    },
     {
       "version": "0.1.0-alpha.728",
       "changes": [

package/dist/heart/daemon/process-manager.js

@@ -672,26 +672,29 @@ class DaemonProcessManager {
             state.snapshot.errorReason = null;
             state.snapshot.fixHint = null;
         }
-        const crashed = !state.stopRequested && code !== 0;
+        const unexpectedExit = !state.stopRequested;
+        const crashed = unexpectedExit && code !== 0;
         const now = this.currentTimeMs();
         const startedAt = state.snapshot.startedAt ? Date.parse(state.snapshot.startedAt) : now;
         const runDuration = Math.max(0, now - startedAt);
         (0, runtime_1.emitNervesEvent)({
-            level: crashed ? "warn" : "info",
+            level: unexpectedExit ? "warn" : "info",
             component: "daemon",
             event: "daemon.agent_exit",
             message: "managed agent process exited",
-            meta: { agent: state.config.name, code, signal, crashed, runDurationMs: runDuration },
+            meta: { agent: state.config.name, code, signal, crashed, unexpectedExit, runDurationMs: runDuration },
         });
-        if (!crashed) {
+        if (runDuration >= this.stabilityThresholdMs) {
+            state.snapshot.backoffMs = this.initialBackoffMs;
+        }
+        if (!unexpectedExit) {
             state.snapshot.status = "stopped";
-            if (runDuration >= this.stabilityThresholdMs) {
-                state.snapshot.backoffMs = this.initialBackoffMs;
-            }
             this.notifySnapshotChange(state.snapshot);
             return;
         }
-        state.snapshot.lastCrashAt = new Date(now).toISOString();
+        if (crashed) {
+            state.snapshot.lastCrashAt = new Date(now).toISOString();
+        }
         if (state.config.autoStart === false) {
             state.snapshot.status = "stopped";
             state.snapshot.errorReason = "non-autostart worker exited unexpectedly; not respawning";
@@ -699,46 +702,45 @@ class DaemonProcessManager {
             this.notifySnapshotChange(state.snapshot);
             return;
         }
-        // Fast-crash detection: if the agent dies within 5 seconds of starting, it's likely
-        // a configuration issue (missing credentials, bad provider, etc.) not a transient failure.
-        // After 3 consecutive fast crashes, stop retrying and mark as config-failed.
+        // Fast-exit detection: repeated early exits, including clean exit-code-0
+        // terminations, indicate a startup/configuration or lifecycle fault rather
+        // than a stable long-lived worker. Bound them without claiming a crash.
         const FAST_CRASH_THRESHOLD_MS = 5000;
         const FAST_CRASH_MAX = 3;
         if (runDuration < FAST_CRASH_THRESHOLD_MS) {
             state.fastCrashCount = state.fastCrashCount + 1;
             if (state.fastCrashCount >= FAST_CRASH_MAX) {
                 state.snapshot.status = "crashed";
-                // Capture the fast-crash diagnosis on the snapshot so it surfaces
-                // via the pulse. The error message is prescriptive: it tells the
-                // user (and their sibling agents) exactly what to do.
-                state.snapshot.errorReason = `agent crashed ${FAST_CRASH_MAX} times within ${FAST_CRASH_THRESHOLD_MS}ms of startup — likely a configuration issue (missing credentials, bad provider).`;
-                state.snapshot.fixHint = `Fix the config and run \`ouro up\` to restart, or check daemon logs for the underlying error.`;
+                // Capture the fast-exit diagnosis on the snapshot so it surfaces via
+                // the pulse with a prescriptive, cause-neutral repair path.
+                state.snapshot.errorReason = `agent exited unexpectedly ${FAST_CRASH_MAX} times within ${FAST_CRASH_THRESHOLD_MS}ms of startup — likely a startup/configuration or lifecycle issue.`;
+                state.snapshot.fixHint = `Inspect daemon logs and lifecycle/configuration, then run \`ouro up\` to restart.`;
                 (0, runtime_1.emitNervesEvent)({
                     level: "error",
                     component: "daemon",
                     event: "daemon.agent_config_failure",
-                    message: `agent crashed ${FAST_CRASH_MAX} times within ${FAST_CRASH_THRESHOLD_MS}ms of startup — likely a configuration issue (missing credentials, bad provider). Not retrying. Fix the config and run \`ouro up\` to restart.`,
-                    meta: { agent: state.config.name, fastCrashCount: state.fastCrashCount, avgRunDurationMs: runDuration },
+                    message: `agent exited unexpectedly ${FAST_CRASH_MAX} times within ${FAST_CRASH_THRESHOLD_MS}ms of startup. Not retrying. Inspect lifecycle/configuration and run \`ouro up\` to restart.`,
+                    meta: { agent: state.config.name, fastUnexpectedExitCount: state.fastCrashCount, avgRunDurationMs: runDuration },
                 });
                 this.notifySnapshotChange(state.snapshot);
                 return; // Don't schedule cooldown recovery — this needs human/agent intervention
             }
         }
         else {
-            // Reset fast-crash counter on a stable run
+            // Reset the fast unexpected-exit counter on a stable run.
             state.fastCrashCount = 0;
         }
         state.crashTimestamps = state.crashTimestamps.filter((crashTs) => crashTs >= startOfHour(now));
         state.crashTimestamps.push(now);
         if (state.crashTimestamps.length > this.maxRestartsPerHour) {
             state.snapshot.status = "crashed";
-            state.snapshot.errorReason = `agent exceeded restart limit (${this.maxRestartsPerHour}/hr) — entering cooldown`;
-            state.snapshot.fixHint = "investigate why the agent keeps crashing; cooldown will retry shortly";
+            state.snapshot.errorReason = `agent exceeded unexpected-exit restart limit (${this.maxRestartsPerHour}/hr) — entering cooldown`;
+            state.snapshot.fixHint = "investigate why the agent keeps exiting unexpectedly; cooldown will retry shortly";
             (0, runtime_1.emitNervesEvent)({
                 level: "error",
                 component: "daemon",
                 event: "daemon.agent_restart_exhausted",
-                message: "managed agent exceeded restart limit and is marked crashed",
+                message: "managed agent exceeded unexpected-exit restart limit and is marked crashed",
                 meta: { agent: state.config.name, maxRestartsPerHour: this.maxRestartsPerHour },
             });
             this.notifySnapshotChange(state.snapshot);

package/dist/heart/daemon/sense-manager.js

@@ -488,7 +488,7 @@ function readBlueBubblesRuntimeFacts(agent, bundlesRoot, snapshot) {
             ...proofFacts,
             lastFailure: "listener process is not running",
             failureLayer: "listener",
-            recoveryAction: "daemon health monitor will restart the BlueBubbles listener when its probe fails",
+            recoveryAction: "daemon process manager will restart the BlueBubbles listener after an unsolicited exit",
         };
     }
     if (state.upstreamStatus === "error") {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@ouro.bot/cli",
-  "version": "0.1.0-alpha.728",
+  "version": "0.1.0-alpha.729",
   "engines": {
     "node": ">=22"
   },