prizmkit 1.1.66 → 1.1.68

package/bundled/dev-pipeline/scripts/parse-stream-progress.py

@@ -23,6 +23,7 @@ import tempfile
 import time
 from collections import Counter
 from datetime import datetime, timezone
+from pathlib import Path
 
 
 # Ordered pipeline phases — index defines forward-only progression.
@@ -76,6 +77,13 @@ class ProgressTracker:
         self.event_format = ""
         self.active_subagent_count = 0
         self.subagent_status_counts = Counter()
+        self.codex_child_thread_ids = set()
+        self.child_session_files = []
+        self.child_total_bytes = 0
+        self.child_activity_signature = ""
+        self.last_child_activity_at = ""
+        self._codex_child_session_paths = {}
+        self._last_child_scan_at = 0.0
         self._text_buffer = ""
         self._in_tool_use = False
         self._current_tool_input_parts = []
@@ -113,6 +121,9 @@ class ProgressTracker:
 
                 elif item_type == "collab_tool_call":
                     tool_name = item.get("tool", "collab")
+                    self._record_codex_child_thread_ids(
+                        item.get("receiver_thread_ids")
+                    )
                     if event_type == "item.started":
                         self.current_tool = tool_name
                         self.tool_call_counts[tool_name] += 1
@@ -345,8 +356,117 @@ class ProgressTracker:
         self.subagent_status_counts = counts
         self.active_subagent_count = active
 
+    def _record_codex_child_thread_ids(self, thread_ids):
+        """Remember Codex child thread IDs reported by collab tool calls."""
+        if not isinstance(thread_ids, list):
+            return
+        for thread_id in thread_ids:
+            if isinstance(thread_id, str) and thread_id.strip():
+                self.codex_child_thread_ids.add(thread_id.strip())
+
+    def _codex_sessions_dir(self):
+        """Return the Codex sessions directory for the current environment."""
+        codex_home = os.environ.get("CODEX_HOME")
+        if codex_home:
+            return Path(codex_home).expanduser() / "sessions"
+        return Path.home() / ".codex" / "sessions"
+
+    def _find_codex_child_session_file(self, thread_id):
+        """Find a Codex transcript file for a child thread ID."""
+        sessions_dir = self._codex_sessions_dir()
+        if not sessions_dir.exists():
+            return None
+
+        try:
+            matches = list(sessions_dir.rglob(f"*{thread_id}.jsonl"))
+        except OSError:
+            return None
+
+        if not matches:
+            return None
+
+        try:
+            matches.sort(key=lambda path: path.stat().st_mtime, reverse=True)
+        except OSError:
+            pass
+        return str(matches[0])
+
+    def refresh_child_session_activity(self, force=False):
+        """Refresh Codex child transcript file stats.
+
+        The heartbeat monitor uses this activity signature to treat subagent
+        transcript growth as real progress while the parent Codex session is
+        blocked in `wait`.
+        """
+        previous_signature = self.child_activity_signature
+
+        if not self.codex_child_thread_ids:
+            self.child_session_files = []
+            self.child_total_bytes = 0
+            self.child_activity_signature = ""
+            self.last_child_activity_at = ""
+            return previous_signature != self.child_activity_signature
+
+        now = time.monotonic()
+        should_scan = (
+            force
+            or self._last_child_scan_at == 0.0
+            or (now - self._last_child_scan_at >= 2.0)
+        )
+        if should_scan:
+            for thread_id in sorted(self.codex_child_thread_ids):
+                path = self._codex_child_session_paths.get(thread_id)
+                if not path or not os.path.exists(path):
+                    found = self._find_codex_child_session_file(thread_id)
+                    if found:
+                        self._codex_child_session_paths[thread_id] = found
+            self._last_child_scan_at = now
+
+        files = []
+        signature_parts = []
+        total_bytes = 0
+        max_mtime = 0.0
+
+        for thread_id in sorted(self.codex_child_thread_ids):
+            path = self._codex_child_session_paths.get(thread_id)
+            if not path:
+                continue
+            try:
+                stat = os.stat(path)
+            except OSError:
+                continue
+
+            total_bytes += stat.st_size
+            max_mtime = max(max_mtime, stat.st_mtime)
+            signature_parts.append(
+                f"{thread_id}:{stat.st_size}:{getattr(stat, 'st_mtime_ns', int(stat.st_mtime * 1_000_000_000))}"
+            )
+            files.append(
+                {
+                    "thread_id": thread_id,
+                    "path": path,
+                    "size": stat.st_size,
+                    "mtime": datetime.fromtimestamp(
+                        stat.st_mtime, timezone.utc
+                    ).strftime("%Y-%m-%dT%H:%M:%SZ"),
+                }
+            )
+
+        self.child_session_files = files
+        self.child_total_bytes = total_bytes
+        self.child_activity_signature = "|".join(signature_parts)
+        self.last_child_activity_at = (
+            datetime.fromtimestamp(max_mtime, timezone.utc).strftime(
+                "%Y-%m-%dT%H:%M:%SZ"
+            )
+            if max_mtime
+            else ""
+        )
+        return previous_signature != self.child_activity_signature
+
     def to_dict(self):
         """Export current state as a dictionary for JSON serialization."""
+        self.refresh_child_session_activity()
         tool_calls = [
             {"name": name, "count": count}
             for name, count in self.tool_call_counts.most_common()
@@ -367,6 +487,11 @@ class ProgressTracker:
             "total_tool_calls": self.total_tool_calls,
             "active_subagent_count": self.active_subagent_count,
             "subagent_states": subagent_states,
+            "child_thread_ids": sorted(self.codex_child_thread_ids),
+            "child_session_files": self.child_session_files,
+            "child_total_bytes": self.child_total_bytes,
+            "child_activity_signature": self.child_activity_signature,
+            "last_child_activity_at": self.last_child_activity_at,
             "last_text_snippet": self.last_text_snippet,
             "is_active": self.is_active,
             "errors": self.errors[-10:],  # Keep last 10 errors
@@ -397,6 +522,15 @@ def tail_and_parse(session_log, progress_file, poll_interval=0.5):
     tracker = ProgressTracker()
     last_write_state = None
 
+    def state_key(state):
+        return (
+            state["message_count"],
+            state["current_tool"],
+            state["current_phase"],
+            state["total_tool_calls"],
+            state.get("child_activity_signature", ""),
+        )
+
     # Wait for log file to appear
     wait_count = 0
     while not os.path.exists(session_log):
@@ -428,22 +562,20 @@ def tail_and_parse(session_log, progress_file, poll_interval=0.5):
 
                 # Write progress if state changed
                 current_state = tracker.to_dict()
-                state_key = (
-                    current_state["message_count"],
-                    current_state["current_tool"],
-                    current_state["current_phase"],
-                    current_state["total_tool_calls"],
-                )
-                if state_key != last_write_state:
+                current_state_key = state_key(current_state)
+                if current_state_key != last_write_state:
                     atomic_write_json(current_state, progress_file)
-                    last_write_state = state_key
+                    last_write_state = current_state_key
             else:
                 idle_count += 1
-                # After 2 seconds of no new data, write current state anyway
-                # (ensures progress.json stays fresh)
-                if idle_count == 4:
+                # Every 2 seconds of no parent log data, refresh child Codex
+                # transcript stats and write if child activity advanced.
+                if idle_count % 4 == 0:
                     current_state = tracker.to_dict()
-                    atomic_write_json(current_state, progress_file)
+                    current_state_key = state_key(current_state)
+                    if current_state_key != last_write_state or idle_count == 4:
+                        atomic_write_json(current_state, progress_file)
+                        last_write_state = current_state_key
 
                 # After 3600 idle cycles (30 min), mark inactive and exit
                 if idle_count > 3600:

package/bundled/dev-pipeline/scripts/update-bug-status.py

@@ -41,6 +41,7 @@ SESSION_STATUS_VALUES = [
     "failed",
     "crashed",
     "timed_out",
+    "infra_error",
     "commit_missing",
     "docs_missing",
     "merge_conflict",
@@ -280,6 +281,16 @@ def action_update(args, bug_list_path, state_dir):
         bs["sessions"] = []
         bs["last_session_id"] = None
 
+        err = update_bug_in_list(bug_list_path, bug_id, new_status)
+        if err:
+            error_out("Failed to update .prizmkit/plans/bug-fix-list.json: {}".format(err))
+            return
+    elif session_status == "infra_error":
+        new_status = "pending"
+        bs["infra_error_count"] = bs.get("infra_error_count", 0) + 1
+        bs["last_infra_error_session_id"] = session_id
+        bs["resume_from_phase"] = None
+
         err = update_bug_in_list(bug_list_path, bug_id, new_status)
         if err:
             error_out("Failed to update .prizmkit/plans/bug-fix-list.json: {}".format(err))
@@ -333,6 +344,10 @@ def action_update(args, bug_list_path, state_dir):
     if session_status in ("commit_missing", "docs_missing", "merge_conflict"):
         summary["degraded_reason"] = session_status
         summary["restart_policy"] = "finalization_retry"
+    elif session_status == "infra_error":
+        summary["restart_policy"] = "infra_retry"
+        summary["infra_error_count"] = bs.get("infra_error_count", 0)
+        summary["artifacts_preserved"] = True
     elif session_status != "success":
         summary["restart_policy"] = "full_restart"
         summary["cleanup_performed"] = cleaned

package/bundled/dev-pipeline/scripts/update-feature-status.py

@@ -45,6 +45,7 @@ SESSION_STATUS_VALUES = [
     "failed",
     "crashed",
     "timed_out",
+    "infra_error",
     "commit_missing",
     "docs_missing",
     "merge_conflict",
@@ -645,6 +646,19 @@ def action_update(args, feature_list_path, state_dir):
         fs["sessions"] = []
         fs["last_session_id"] = None
 
+        err = update_feature_in_list(feature_list_path, feature_id, new_status)
+        if err:
+            error_out("Failed to update .prizmkit/plans/feature-list.json: {}".format(err))
+            return
+    elif session_status == "infra_error":
+        # AI CLI/provider outage, auth failure, gateway error, etc.
+        # This is outside the code's control, so keep the item pending without
+        # consuming the task's retry budget.
+        new_status = "pending"
+        fs["infra_error_count"] = fs.get("infra_error_count", 0) + 1
+        fs["last_infra_error_session_id"] = session_id
+        fs["resume_from_phase"] = None
+
         err = update_feature_in_list(feature_list_path, feature_id, new_status)
         if err:
             error_out("Failed to update .prizmkit/plans/feature-list.json: {}".format(err))
@@ -701,6 +715,10 @@ def action_update(args, feature_list_path, state_dir):
     if session_status in ("commit_missing", "docs_missing", "merge_conflict"):
         summary["degraded_reason"] = session_status
         summary["restart_policy"] = "finalization_retry"
+    elif session_status == "infra_error":
+        summary["restart_policy"] = "infra_retry"
+        summary["infra_error_count"] = fs.get("infra_error_count", 0)
+        summary["artifacts_preserved"] = True
     elif session_status != "success":
         summary["restart_policy"] = "preserve_and_retry"
         summary["artifacts_preserved"] = True

package/bundled/dev-pipeline/scripts/update-refactor-status.py

@@ -42,6 +42,7 @@ SESSION_STATUS_VALUES = [
     "failed",
     "crashed",
     "timed_out",
+    "infra_error",
     "commit_missing",
     "docs_missing",
     "merge_conflict",
@@ -314,6 +315,16 @@ def action_update(args, refactor_list_path, state_dir):
         rs["sessions"] = []
         rs["last_session_id"] = None
 
+        err = update_refactor_in_list(refactor_list_path, refactor_id, new_status)
+        if err:
+            error_out("Failed to update .prizmkit/plans/refactor-list.json: {}".format(err))
+            return
+    elif session_status == "infra_error":
+        new_status = "pending"
+        rs["infra_error_count"] = rs.get("infra_error_count", 0) + 1
+        rs["last_infra_error_session_id"] = session_id
+        rs["resume_from_phase"] = None
+
         err = update_refactor_in_list(refactor_list_path, refactor_id, new_status)
         if err:
             error_out("Failed to update .prizmkit/plans/refactor-list.json: {}".format(err))
@@ -376,6 +387,10 @@ def action_update(args, refactor_list_path, state_dir):
     if session_status in ("commit_missing", "docs_missing", "merge_conflict"):
         summary["degraded_reason"] = session_status
         summary["restart_policy"] = "finalization_retry"
+    elif session_status == "infra_error":
+        summary["restart_policy"] = "infra_retry"
+        summary["infra_error_count"] = rs.get("infra_error_count", 0)
+        summary["artifacts_preserved"] = True
     elif session_status != "success":
         summary["restart_policy"] = "full_restart"
         summary["cleanup_performed"] = cleaned

package/bundled/dev-pipeline/tests/test_auto_skip.py

@@ -303,6 +303,45 @@ def _run_get_next(fl_path, state_dir):
     return result.stdout.strip()
 
 
+def _run_update(fl_path, state_dir, feature_id, session_status, session_id="session-1", max_retries=3):
+    cmd = [
+        "python3", _SCRIPT,
+        "--feature-list", fl_path,
+        "--state-dir", state_dir,
+        "--feature-id", feature_id,
+        "--session-status", session_status,
+        "--session-id", session_id,
+        "--max-retries", str(max_retries),
+        "--action", "update",
+    ]
+    result = subprocess.run(cmd, capture_output=True, text=True)
+    assert result.returncode == 0, result.stderr
+    return json.loads(result.stdout)
+
+
+class TestInfraErrorUpdate:
+    def test_infra_error_keeps_pending_without_consuming_retry(self, tmp_path):
+        features = [_make_feature("F-001", "Root", status="in_progress")]
+        fl_path = _write_fl(tmp_path, features)
+        state_dir = _init_state(tmp_path, ["F-001"])
+        status_path = os.path.join(state_dir, "features", "F-001", "status.json")
+        fs = load_feature_status(state_dir, "F-001")
+        fs["retry_count"] = 2
+        write_json_file(status_path, fs)
+
+        result = _run_update(fl_path, state_dir, "F-001", "infra_error", "session-infra", max_retries=3)
+
+        assert result["new_status"] == "pending"
+        assert result["retry_count"] == 2
+        assert result["restart_policy"] == "infra_retry"
+        assert _read_statuses(fl_path)["F-001"] == "pending"
+
+        fs = load_feature_status(state_dir, "F-001")
+        assert fs["retry_count"] == 2
+        assert fs["infra_error_count"] == 1
+        assert fs["last_infra_error_session_id"] == "session-infra"
+
+
 class TestUnskipByFeatureId:
     """Unskip with --feature-id targets a specific failed feature + downstream."""
 

package/bundled/dev-pipeline-windows/.env.example

@@ -22,9 +22,10 @@
 # SESSION_TIMEOUT=0          # Session timeout in seconds (0 = no limit)
 # VERBOSE=1                  # Verbose logging (1=on, 0=off)
 # HEARTBEAT_INTERVAL=30      # Poll interval for session progress/stale checks
-# STALE_KILL_THRESHOLD=900   # Auto-kill session after N seconds without log progress (0 = disabled)
+# STALE_KILL_THRESHOLD=900   # Auto-kill session after N seconds without parent log progress (0 = disabled)
 # STALE_KILL_GRACE_SECONDS=10  # Grace period after stale-kill before force-stopping the job
-# CODEX_SUBAGENT_TIMEOUT_SECONDS=840  # Codex subagent max runtime; defaults to stale threshold - 60
+# CODEX_WAIT_STALE_KILL_THRESHOLD=3600  # Longer no-log window while Codex waits on subagents
+# CODEX_SUBAGENT_TIMEOUT_SECONDS=3300  # Codex subagent max runtime; defaults to wait threshold - 300
 # LOG_CLEANUP_ENABLED=1      # Run periodic session log cleanup
 # LOG_RETENTION_DAYS=14      # Delete session logs older than N days
 # LOG_MAX_TOTAL_MB=1024      # Keep total logs under N MB via oldest-first cleanup

package/bundled/dev-pipeline-windows/SCHEMA_ANALYSIS.md

@@ -346,13 +346,14 @@ pending, in_progress, completed, failed, skipped
 | `SESSION_TIMEOUT` | integer | 0 | 0 = no limit |
 | `VERBOSE` | integer | (not specified) | 1=on, 0=off |
 | `HEARTBEAT_INTERVAL` | integer | 30 | Poll interval for session progress/stale checks |
-| `STALE_KILL_THRESHOLD` | integer | 900 | Auto-kill after N seconds without log progress; 0 disables |
+| `STALE_KILL_THRESHOLD` | integer | 900 | Auto-kill after N seconds without parent log progress; 0 disables |
 | `STALE_KILL_GRACE_SECONDS` | integer | 10 | Grace period after stale-kill before force-stopping |
-| `CODEX_SUBAGENT_TIMEOUT_SECONDS` | integer | 840 | Codex subagent max runtime |
+| `CODEX_WAIT_STALE_KILL_THRESHOLD` | integer | 3600 | Longer no-log stale window while Codex waits on subagents |
+| `CODEX_SUBAGENT_TIMEOUT_SECONDS` | integer | 3300 | Codex subagent max runtime |
 | `LOG_CLEANUP_ENABLED` | boolean | 1 | Periodic session log cleanup |
 | `LOG_RETENTION_DAYS` | integer | 14 | Delete session logs older than N days |
 | `LOG_MAX_TOTAL_MB` | integer | 1024 | Keep total logs under N MB |
-| `STOP_ON_FAILURE` | boolean | 0 | Stop after the first failed task |
+| `STOP_ON_FAILURE` | boolean | 0 | Stop after a task exhausts retries |
 | `ENABLE_DEPLOY` | boolean | 0 | Start deploy session after all tasks complete |
 | `DEV_BRANCH` | string | auto-generated | Optional custom dev branch name |
 | `AUTO_PUSH` | boolean | 0 | Push original branch after successful merge |

package/bundled/dev-pipeline-windows/lib/common.ps1

@@ -145,6 +145,25 @@ function Invoke-PrizmPythonText {
   if ($LASTEXITCODE -ne 0) { throw "Python command failed: $($Arguments -join ' ')" }
 }
 
+function Test-PrizmInfraError {
+  param([string]$SessionLog, [string]$ProgressJson)
+  $parts = @()
+  if ($SessionLog -and (Test-Path $SessionLog)) {
+    try {
+      $text = Get-Content $SessionLog -Raw -ErrorAction Stop
+      if ($text.Length -gt 65536) { $text = $text.Substring($text.Length - 65536) }
+      $parts += $text
+    } catch {}
+  }
+  if ($ProgressJson -and (Test-Path $ProgressJson)) {
+    try { $parts += (Get-Content $ProgressJson -Raw -ErrorAction Stop) } catch {}
+  }
+  if ($parts.Count -eq 0) { return $false }
+
+  $haystack = $parts -join "`n"
+  return ($haystack -match '(?i)auth_unavailable|no auth available|502 Bad Gateway|503 Service Unavailable|504 Gateway Timeout|gateway timeout|upstream (connect )?error|connection reset|ECONNRESET|ETIMEDOUT|ENOTFOUND|EAI_AGAIN|rate limit|rate_limit|temporarily unavailable|overloaded')
+}
+
 function Get-PrizmConfigValue {
   param([string]$ConfigPath, [string]$Key)
   if (-not (Test-Path $ConfigPath)) { return $null }
@@ -345,13 +364,86 @@ function Get-PrizmCodexSubagentTimeoutSeconds {
     return $configuredTimeout
   }
 
-  $outerThreshold = 0
-  $outerThresholdText = if ($env:STALE_KILL_THRESHOLD) { $env:STALE_KILL_THRESHOLD } else { $env:SESSION_TIMEOUT }
-  if ([int]::TryParse($outerThresholdText, [ref]$outerThreshold) -and $outerThreshold -gt 120) {
-    return ($outerThreshold - 60)
+  $waitThreshold = 0
+  $waitThresholdText = if ($env:CODEX_WAIT_STALE_KILL_THRESHOLD) { $env:CODEX_WAIT_STALE_KILL_THRESHOLD } else { '3600' }
+  if ([int]::TryParse($waitThresholdText, [ref]$waitThreshold) -and $waitThreshold -gt 600) {
+    return ($waitThreshold - 300)
   }
 
-  return 840
+  return 3300
+}
+
+function Get-PrizmEffectiveStaleKillThreshold {
+  param(
+    [string]$ProgressFile,
+    [int]$BaseThreshold
+  )
+
+  if ($BaseThreshold -le 0) { return $BaseThreshold }
+  if (-not (Test-Path $ProgressFile)) { return $BaseThreshold }
+
+  try {
+    $progress = Get-Content $ProgressFile -Raw -ErrorAction Stop | ConvertFrom-Json -ErrorAction Stop
+  } catch {
+    return $BaseThreshold
+  }
+
+  $spawnCount = 0
+  if ($progress.event_format -eq 'codex-json' -and $progress.current_tool -eq 'wait' -and $progress.tool_calls) {
+    foreach ($tool in @($progress.tool_calls)) {
+      if ($tool.name -eq 'spawn_agent') {
+        $count = 0
+        if ([int]::TryParse([string]$tool.count, [ref]$count)) { $spawnCount += $count }
+      }
+    }
+  }
+
+  if ($spawnCount -le 0) { return $BaseThreshold }
+
+  $waitThreshold = 0
+  if ([int]::TryParse($env:CODEX_WAIT_STALE_KILL_THRESHOLD, [ref]$waitThreshold) -and $waitThreshold -gt $BaseThreshold) {
+    return $waitThreshold
+  }
+
+  return [Math]::Max($BaseThreshold * 4, 3600)
+}
+
+function Get-PrizmProgressChildActivity {
+  param([string]$ProgressFile)
+
+  $empty = [pscustomobject]@{
+    Signature = ''
+    TotalBytes = 0
+    SessionCount = 0
+  }
+  if (-not (Test-Path $ProgressFile)) { return $empty }
+
+  try {
+    $progress = Get-Content $ProgressFile -Raw -ErrorAction Stop | ConvertFrom-Json -ErrorAction Stop
+  } catch {
+    return $empty
+  }
+
+  $signature = ''
+  if ($progress.PSObject.Properties['child_activity_signature'] -and $progress.child_activity_signature) {
+    $signature = [string]$progress.child_activity_signature
+  }
+
+  $totalBytes = [int64]0
+  if ($progress.PSObject.Properties['child_total_bytes']) {
+    [int64]::TryParse([string]$progress.child_total_bytes, [ref]$totalBytes) | Out-Null
+  }
+
+  $sessionCount = 0
+  if ($progress.PSObject.Properties['child_session_files'] -and $progress.child_session_files) {
+    $sessionCount = @($progress.child_session_files).Count
+  }
+
+  return [pscustomobject]@{
+    Signature = $signature
+    TotalBytes = $totalBytes
+    SessionCount = $sessionCount
+  }
 }
 
 function Test-PrizmCodexJsonSupport {

package/bundled/dev-pipeline-windows/lib/pipeline.ps1

@@ -552,6 +552,7 @@ function Invoke-PrizmPipeline {
     $elapsedSeconds = 0
     $staleSeconds = 0
     $previousLogSize = 0
+    $previousChildActivitySignature = ''
     $wasTimedOut = $false
     $staleKillMarker = Join-Path $logsDir 'stale-kill.json'
     $wasStaleKilled = $false
@@ -568,7 +569,13 @@ function Invoke-PrizmPipeline {
       }
       $growth = $currentLogSize - $previousLogSize
       $previousLogSize = $currentLogSize
-      if ($growth -gt 0) {
+
+      $childActivity = Get-PrizmProgressChildActivity -ProgressFile $progressJson
+      $childSignature = [string]$childActivity.Signature
+      $childAdvanced = ($childSignature -and $childSignature -ne $previousChildActivitySignature)
+      $previousChildActivitySignature = $childSignature
+
+      if ($growth -gt 0 -or $childAdvanced) {
         $staleSeconds = 0
       } else {
         $staleSeconds += $waitSeconds
@@ -580,10 +587,11 @@ function Invoke-PrizmPipeline {
         break
       }
 
-      if ($staleKillThreshold -gt 0 -and $staleSeconds -ge $staleKillThreshold) {
+      $effectiveStaleKillThreshold = Get-PrizmEffectiveStaleKillThreshold -ProgressFile $progressJson -BaseThreshold $staleKillThreshold
+      if ($effectiveStaleKillThreshold -gt 0 -and $staleSeconds -ge $effectiveStaleKillThreshold) {
         $wasStaleKilled = $true
-        Write-PrizmWarn "Session stale-killed (no progress for ${staleKillThreshold}s)"
-        Write-PrizmStaleKillMarker $staleKillMarker $staleSeconds $staleKillThreshold
+        Write-PrizmWarn "Session stale-killed (no progress for ${effectiveStaleKillThreshold}s)"
+        Write-PrizmStaleKillMarker $staleKillMarker $staleSeconds $effectiveStaleKillThreshold
         Stop-PrizmSessionProcess $pidPath
         if ($staleKillGraceSeconds -gt 0) { Start-Sleep -Seconds $staleKillGraceSeconds }
         break
@@ -610,10 +618,16 @@ function Invoke-PrizmPipeline {
     }
     Stop-PrizmProgressParser $parserProcess
 
+    $wasInfraError = ($exitCode -ne 0 -and (Test-PrizmInfraError -SessionLog $sessionLog -ProgressJson $progressJson))
+
     $status = 'crashed'
     if ($wasTimedOut) {
       $status = 'timed_out'
       Write-PrizmWarn "AI session timed out after $timeoutSeconds seconds"
+    } elseif ($wasInfraError) {
+      $status = 'infra_error'
+      Write-PrizmWarn "AI session failed due to AI CLI/provider infrastructure error"
+      Write-PrizmWarn "Infrastructure errors are retried without consuming code retry budget"
     } elseif ($wasStaleKilled -or (Test-Path $staleKillMarker)) {
       Write-PrizmWarn "Session was stale-killed by heartbeat monitor (no progress for too long)"
       Write-PrizmWarn "Stale-killed sessions are treated as failed; dev branch is preserved for inspection"
@@ -637,8 +651,12 @@ function Invoke-PrizmPipeline {
     }
 
     $mergeSucceeded = $true
+    $itemListStatus = ''
     if ($status -eq 'success') {
-      Invoke-PrizmPythonText $python (@((Join-Path $paths.ScriptsDir $updateScript), $listOption, $listPath, '--state-dir', $stateDir, '--action', 'update', $idOption, $CurrentItemId, '--session-id', $sessionId, '--session-status', $status) + $maxRetryArgs)
+      $updateResult = Invoke-PrizmPythonJson $python (@((Join-Path $paths.ScriptsDir $updateScript), $listOption, $listPath, '--state-dir', $stateDir, '--action', 'update', $idOption, $CurrentItemId, '--session-id', $sessionId, '--session-status', $status) + $maxRetryArgs)
+      if ($updateResult -and $updateResult.PSObject.Properties['new_status']) {
+        $itemListStatus = [string]$updateResult.new_status
+      }
 
       if (Test-PrizmGitDirty $paths.ProjectRoot) {
         if ($hadDirtyBaseline) {
@@ -668,7 +686,10 @@ function Invoke-PrizmPipeline {
     }
 
     if ($status -ne 'success') {
-      Invoke-PrizmPythonText $python (@((Join-Path $paths.ScriptsDir $updateScript), $listOption, $listPath, '--state-dir', $stateDir, '--action', 'update', $idOption, $CurrentItemId, '--session-id', $sessionId, '--session-status', $status) + $maxRetryArgs)
+      $updateResult = Invoke-PrizmPythonJson $python (@((Join-Path $paths.ScriptsDir $updateScript), $listOption, $listPath, '--state-dir', $stateDir, '--action', 'update', $idOption, $CurrentItemId, '--session-id', $sessionId, '--session-status', $status) + $maxRetryArgs)
+      if ($updateResult -and $updateResult.PSObject.Properties['new_status']) {
+        $itemListStatus = [string]$updateResult.new_status
+      }
       if ($isGitRepository) {
         Invoke-PrizmGitCommitPath $paths.ProjectRoot $listPath "chore($CurrentItemId): update $idName status" | Out-Null
       }
@@ -679,6 +700,7 @@ function Invoke-PrizmPipeline {
     } else {
       Write-PrizmError "$Kind item failed: $CurrentItemId. Log: $sessionLog"
     }
+    $script:PRIZM_ITEM_LIST_STATUS = $itemListStatus
     $script:PRIZM_ITEM_EXIT_CODE = if ($status -eq 'success' -and $mergeSucceeded) { 0 } else { 1 }
     return
   }
@@ -740,9 +762,11 @@ function Invoke-PrizmPipeline {
       $global:PRIZM_EXIT_CODE = $lastExitCode
       return
     }
-    if ($lastExitCode -ne 0 -and $stopOnFailure) {
+    if ($lastExitCode -ne 0 -and $stopOnFailure -and $script:PRIZM_ITEM_LIST_STATUS -eq 'failed') {
       $global:PRIZM_EXIT_CODE = $lastExitCode
       return
+    } elseif ($lastExitCode -ne 0 -and $stopOnFailure) {
+      Write-PrizmInfo "STOP_ON_FAILURE: $nextItemId is $($script:PRIZM_ITEM_LIST_STATUS); retry budget not exhausted, continuing."
     }
   }
 }

package/bundled/dev-pipeline-windows/run-recovery.ps1

@@ -110,6 +110,7 @@ $job = Start-Job -ScriptBlock {
 $elapsedSeconds = 0
 $staleSeconds = 0
 $previousLogSize = 0
+$previousChildActivitySignature = ''
 $wasTimedOut = $false
 $wasStaleKilled = $false
 while ($true) {
@@ -123,7 +124,13 @@ while ($true) {
   if (Test-Path $logPath) { $currentLogSize = [int64](Get-Item $logPath).Length }
   $growth = $currentLogSize - $previousLogSize
   $previousLogSize = $currentLogSize
-  if ($growth -gt 0) { $staleSeconds = 0 } else { $staleSeconds += $waitSeconds }
+
+  $childActivity = Get-PrizmProgressChildActivity -ProgressFile $progressPath
+  $childSignature = [string]$childActivity.Signature
+  $childAdvanced = ($childSignature -and $childSignature -ne $previousChildActivitySignature)
+  $previousChildActivitySignature = $childSignature
+
+  if ($growth -gt 0 -or $childAdvanced) { $staleSeconds = 0 } else { $staleSeconds += $waitSeconds }
 
   if ($timeoutSeconds -gt 0 -and $elapsedSeconds -ge $timeoutSeconds) {
     $wasTimedOut = $true