opencode-auto-resume 1.1.19 → 1.1.20

package/README.md

@@ -10,7 +10,7 @@ LLM sessions fail in predictable ways. This plugin monitors all sessions and aut
 
 ### Stall recovery
 
-The stream goes silent but the session stays "busy". The UI shows a blinking cursor with no progress. If no events arrive for 48 seconds (`chunkTimeoutMs` + `gracePeriodMs`), the plugin sends `"continue"` with exponential backoff. After 3 failed attempts it gives up.
+The stream goes silent but the session stays "busy". The UI shows a blinking cursor with no progress. If no events arrive for ~3 minutes (`chunkTimeoutMs` + `gracePeriodMs`), the plugin sends `"continue"` with exponential backoff. After 3 failed attempts it gives up.
 
 The `busyStallStrategy` option controls this path: `"continue"` (default), `"abort"` (abort-first), or `"off"` — see [Recovery model](#recovery-model) for what a busy prompt can and cannot do while the runner is live.
 
@@ -250,7 +250,7 @@ If the assistant claims the task is done ("task done", "finished", "all complete
 The session is not stalled while the ball is in the user's court. Two gates stand down idle nudges:
 
 - **Awaiting input**: the newest assistant message holds a `tool_use` part with `state.status: "pending"` (e.g. an open `question` tool call). All idle checks, the periodic recheck, delayed action-intent callbacks, and the tool-text scan skip prompting until a newer user message clears the gate. Completed tool calls never engage it.
-- **Recently active user**: any inbound user message within `activeUserWindowMs` (default 15 minutes) means the user is engaged — likely composing a reply, which leaves no pending tool call behind. Open-todos nudges (idle + periodic), the tool-text reminder fallback, and action-intent callbacks stand down until the window expires.
+- **Recently active user**: any inbound user message within `activeUserWindowMs` (default 5 minutes) means the user is engaged — likely composing a reply, which leaves no pending tool call behind. Open-todos nudges (idle + periodic), the tool-text reminder fallback, and action-intent callbacks stand down until the window expires.
 
 ---
 
@@ -367,7 +367,7 @@ With options:
 {
   "plugin": [
     ["opencode-auto-resume", {
-      "chunkTimeoutMs": 45000,
+      "chunkTimeoutMs": 180000,
       "gracePeriodMs": 3000,
       "maxRetries": 3
     }]
@@ -382,7 +382,7 @@ With options:
   "plugin": [
     [
       "file:///home/YOURUSER/.config/opencode/plugins/opencode-auto-resume/dist/index.js",
-      { "chunkTimeoutMs": 45000, "maxRetries": 3 }
+      { "chunkTimeoutMs": 180000, "maxRetries": 3 }
     ]
   ]
 }
@@ -392,7 +392,7 @@ With options:
 
 | Option | Default | Description |
 |---|---|---|
-| `chunkTimeoutMs` | `45000` | Inactivity timeout before considering stream stalled |
+| `chunkTimeoutMs` | `180000` | Inactivity timeout before considering stream stalled |
 | `gracePeriodMs` | `3000` | Extra wait before acting (lets ESC/status events arrive) |
 | `checkIntervalMs` | `5000` | Timer poll interval |
 | `maxRetries` | `3` | Max auto-resume attempts before giving up |
@@ -407,6 +407,7 @@ With options:
 | `toolTextCheckDelayMs` | `3000` | Delay before scanning an idle session for tool-as-text; also the recovery watchdog delay |
 | `minActivityGapMs` | `1000` | Skip recovery if the session was active within this gap |
 | `warmupMs` | `15000` | Action-intent detection disabled while a session is younger than this |
+| `discoveryDelayMs` | `5000` | Delay before the initial session discovery run after plugin attach |
 | `debug` | `false` | Enable `[debug]` console diagnostics |
 | `resumeOnActionIntent` | `true` | Enable action-intent (`:`-terminated line) nudges |
 | `continuePrompt` | `"continue"` | Prompt text for stall/streaming/dead-stream recovery |
@@ -420,7 +421,7 @@ With options:
 | `silentDeadStreamMinTokens` | `200` | Min output tokens to treat a textless `finish:"unknown"` message as a dead stream |
 | `busyStallStrategy` | `"continue"` | Busy-stall response: `"continue"`, `"abort"` (abort-first), or `"off"` (disabled) |
 | `contextSaturationThreshold` | `0.85` | Ratio of used/usable context that routes a saturated parent to magic-context `ctx-wrapup` (only when magic-context is installed) |
-| `activeUserWindowMs` | `900000` | Inbound-user-message recency window (15 min) during which idle nudges stand down (user likely composing) |
+| `activeUserWindowMs` | `300000` | Inbound-user-message recency window (5 min) during which idle nudges stand down (user likely composing) |
 | `subagentNativeCompactionEnabled` | `false` | Opt-in native `session.summarize()` for saturated subagent sessions (no magic-context detection required) |
 
 Message patterns are matched case-insensitively. Error names use exact match.
@@ -445,7 +446,7 @@ The plugin handles all recovery automatically — no manual intervention needed.
 | Problem | Solution |
 |---|---|
 | Resumes after ESC | Increase `gracePeriodMs` to `5000` |
-| Too aggressive | Increase `chunkTimeoutMs` to `60000` |
+| Stall recovery too aggressive | Decrease `chunkTimeoutMs` (e.g. `45000`) |
 | Too slow to react | Decrease `checkIntervalMs` to `2000` |
 | Orphan parent not detected | Increase `subagentWaitMs` to `20000` |
 | Hallucination loop not caught | Decrease `loopMaxContinues` to `2` |

package/dist/index.js

@@ -12335,9 +12335,10 @@ function tool(input) {
 }
 tool.schema = exports_external;
 // src/index.ts
-var DEFAULT_CHUNK_TIMEOUT_MS = 45000;
+var DEFAULT_CHUNK_TIMEOUT_MS = 180000;
 var DEFAULT_CHECK_INTERVAL_MS = 5000;
-var DEFAULT_ACTIVE_USER_WINDOW_MS = 15 * 60000;
+var DEFAULT_DISCOVERY_DELAY_MS = 5000;
+var DEFAULT_ACTIVE_USER_WINDOW_MS = 5 * 60000;
 var DEFAULT_GRACE_PERIOD_MS = 3000;
 var DEFAULT_MAX_RETRIES = 3;
 var DEFAULT_MAX_BACKOFF_MS = 8000;
@@ -12352,6 +12353,7 @@ var DEFAULT_MIN_ACTIVITY_GAP_MS = 1000;
 var DEFAULT_WARMUP_MS = 15000;
 var DEFAULT_SILENT_DEAD_STREAM_MIN_TOKENS = 200;
 var DEFAULT_DEBUG = false;
+var OOC_ERROR_RE = /exceeds the available context size|context size \(\d+\)|too large to compact|too many tokens|prompt is too long/i;
 var DEFAULT_STREAMING_FAILURE_ERROR_NAMES = [
   "ProviderError",
   "APIError",
@@ -12659,6 +12661,7 @@ var AutoResumePlugin = async (ctx, options) => {
   const maxRecoveryRetries = options?.maxRecoveryRetries ?? DEFAULT_MAX_RECOVERY_RETRIES;
   const minActivityGapMs = options?.minActivityGapMs ?? DEFAULT_MIN_ACTIVITY_GAP_MS;
   const warmupMs = options?.warmupMs ?? DEFAULT_WARMUP_MS;
+  const discoveryDelayMs = options?.discoveryDelayMs ?? DEFAULT_DISCOVERY_DELAY_MS;
   const debug = options?.debug ?? DEFAULT_DEBUG;
   const streamingFailureErrorNames = options?.streamingFailureErrorNames ?? DEFAULT_STREAMING_FAILURE_ERROR_NAMES;
   const streamingFailureMessagePatterns = options?.streamingFailureMessagePatterns ?? DEFAULT_STREAMING_FAILURE_MESSAGE_PATTERNS;
@@ -12770,6 +12773,8 @@ var AutoResumePlugin = async (ctx, options) => {
         resumeAttempts: 0,
         lastRetryAt: 0,
         gaveUp: false,
+        oocLocked: false,
+        oocLockReason: null,
         orphanWatchStartAt: null,
         aborting: false,
         pluginAbortInFlight: false,
@@ -12946,6 +12951,10 @@ var AutoResumePlugin = async (ctx, options) => {
       await log("debug", `${short(sid)} - gaveUp latched, refusing further continue prompts`);
       return;
     }
+    if (w.oocLocked) {
+      await log("debug", `${short(sid)} - oocLocked latched, refusing further continue prompts (reason: ${w.oocLockReason ?? "n/a"})`);
+      return;
+    }
     if (!w.continuing)
       dbg(`State transition on ${short(sid)}: continuing=false -> true`);
     w.continuing = true;
@@ -13831,14 +13840,13 @@ var AutoResumePlugin = async (ctx, options) => {
           await log("info", `${short(sid)} - max open-todos nudges (${maxRetries}) reached, waiting for activity`);
           return;
         }
-        w.todoNudgeAttempts++;
       } else if (isDoneClaimNoTodos) {
         w.doneClaimNoTodosAttempts++;
       } else {
         w.toolTextRecovered = true;
         w.toolTextAttempts++;
       }
-      const attemptNum = isOpenTodosReminder ? w.todoNudgeAttempts : isDoneClaimNoTodos ? w.doneClaimNoTodosAttempts : w.toolTextAttempts;
+      const attemptNum = isOpenTodosReminder ? w.todoNudgeAttempts + 1 : isDoneClaimNoTodos ? w.doneClaimNoTodosAttempts : w.toolTextAttempts;
       await log("info", `${bestCandidate.source} detected on ${short(sid)}! ` + `Attempt ${attemptNum}/${maxRetries}. Sending recovery prompt...`);
       const timeSinceActivity = Date.now() - w.lastActivityAt;
       if (timeSinceActivity < minActivityGapMs) {
@@ -13884,6 +13892,8 @@ var AutoResumePlugin = async (ctx, options) => {
       return false;
     if (w.aborting)
       return false;
+    if (w.oocLocked)
+      return false;
     const idleSec = Math.round((Date.now() - (w.orphanWatchStartAt ?? w.lastActivityAt)) / 1000);
     await log("info", `Abort+Resume on ${short(sid)} (${idleSec}s idle). Aborting...`);
     w.pluginAbortInFlight = true;
@@ -13968,12 +13978,13 @@ var AutoResumePlugin = async (ctx, options) => {
     try {
       const response = await ctx.client.session.list();
       const list = extractMessages(response);
+      const statusMap = await getSessionStatusMap();
       for (const s of list) {
         const sid = s.id;
         if (sid && typeof sid === "string" && sid.startsWith("ses_")) {
           const isNew = !sessions.has(sid);
           ensureWatch(sid);
-          const status = s.status;
+          const status = s.status ?? statusMap[sid];
           if (status) {
             const w = sessions.get(sid);
             w.status = status;
@@ -14216,7 +14227,7 @@ var AutoResumePlugin = async (ctx, options) => {
     }, SESSION_DISCOVERY_INTERVAL_MS);
     if (discoveryTimer.unref)
       discoveryTimer.unref();
-    setTimeout(discoverSessions, 5000);
+    setTimeout(discoverSessions, discoveryDelayMs);
   }
   startTimer();
   async function handleEvent(ev) {
@@ -14237,7 +14248,11 @@ var AutoResumePlugin = async (ctx, options) => {
           if (w.pendingRecovery) {
             dbg(`Pending recovery cleared on ${short(sid)}: reason=session-busy`);
           }
-          resetBusyFlags(w);
+          if (w.continuing) {
+            dbg(`busy on ${short(sid)} via recovery prompt - preserving resumeAttempts=${w.resumeAttempts} gaveUp=${w.gaveUp}`);
+          } else {
+            resetBusyFlags(w);
+          }
           prevBusyCount = busyCount();
           log("debug", `${short(sid)} -> busy (${prevBusyCount})`);
         } else if (statusType === "interrupted") {
@@ -14379,12 +14394,14 @@ var AutoResumePlugin = async (ctx, options) => {
               if (isCelebration) {
                 await log("info", `${short(sid)} - \uD83C\uDF89 detected but ${open.length} open todos remain, sending nudge`);
                 const reminder = buildOpenTodosReminder(todos);
-                await tryResume(sid, w, "Idle with open todos (celebration false positive)", reminder);
-                w.todoNudgeAttempts++;
+                const sent = await tryResume(sid, w, "Idle with open todos (celebration false positive)", reminder);
+                if (sent)
+                  w.todoNudgeAttempts++;
               } else {
                 const reminder = buildOpenTodosReminder(todos);
-                await tryResume(sid, w, "Idle with open todos", reminder);
-                w.todoNudgeAttempts++;
+                const sent = await tryResume(sid, w, "Idle with open todos", reminder);
+                if (sent)
+                  w.todoNudgeAttempts++;
               }
             }
           }
@@ -14626,6 +14643,11 @@ var AutoResumePlugin = async (ctx, options) => {
           if (w) {
             w.pendingTools = 0;
             w.pendingCommands = 0;
+            if (!w.oocLocked && OOC_ERROR_RE.test(errorMessage ?? "")) {
+              w.oocLocked = true;
+              w.oocLockReason = (errorMessage ?? "").slice(0, 200);
+              await log("warn", `${short(sid)} - OOC error latched, halting recovery (reason: ${w.oocLockReason})`);
+            }
           }
         }
         break;
@@ -14638,7 +14660,11 @@ var AutoResumePlugin = async (ctx, options) => {
           if (w.pendingRecovery) {
             dbg(`Pending recovery cleared on ${short(sid)}: reason=user-command`);
           }
-          resetBusyFlags(w);
+          if (w.continuing) {
+            dbg(`command on ${short(sid)} via recovery prompt - preserving resumeAttempts=${w.resumeAttempts} gaveUp=${w.gaveUp}`);
+          } else {
+            resetBusyFlags(w);
+          }
           w.pendingCommands = Math.max(0, w.pendingCommands - 1);
           w.lastActivityAt = Date.now();
         }
@@ -14726,6 +14752,11 @@ ${reminder}`;
       w.lastActivityAt = Date.now();
       if (w.continuing)
         return;
+      if (w.oocLocked) {
+        w.oocLocked = false;
+        w.oocLockReason = null;
+        await log("info", `${short(sid)} - genuine user message, clearing oocLocked`);
+      }
       if (w.userCancelled || w.completionSignaled) {
         w.userCancelled = false;
         w.completionSignaled = false;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "opencode-auto-resume",
-  "version": "1.1.19",
+  "version": "1.1.20",
   "description": "OpenCode plugin that automatically resumes stalled LLM sessions when thinking/streaming freezes mid-generation.",
   "keywords": [
     "opencode",