From 6c9b18d7b261dfbac4fa5502d25b5a8afa07dc7a Mon Sep 17 00:00:00 2001
From: A <258483684+la14-1@users.noreply.github.com>
Date: Tue, 10 Feb 2026 17:49:54 -0800
Subject: [PATCH] fix: Add monitoring loop instructions and increase watchdog
 to 10 min (#315)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The team lead was spawning 6 agents then exiting because the prompt
lacked explicit instructions to stay alive and wait for messages (the
discovery prompt has this, refactor didn't). Added the WaitForMessage
monitoring loop pattern from discovery.sh.

Also increased IDLE_TIMEOUT from 180s to 600s — 3 min was too
aggressive, killing legitimate cycles where agents are working and the
leader is waiting for their responses.

Co-authored-by: A <6723574+louisgv@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
---
 .../skills/setup-trigger-service/refactor.sh  | 29 +++++++++++++++++--
 1 file changed, 27 insertions(+), 2 deletions(-)

diff --git a/.claude/skills/setup-trigger-service/refactor.sh b/.claude/skills/setup-trigger-service/refactor.sh
index 2974751d..c380ad3d 100755
--- a/.claude/skills/setup-trigger-service/refactor.sh
+++ b/.claude/skills/setup-trigger-service/refactor.sh
@@ -417,7 +417,7 @@ git worktree remove WORKTREE_BASE_PLACEHOLDER/BRANCH-NAME
 7. Community-coordinator engages issues FIRST — posts acknowledgments before other agents start investigating
 8. Community-coordinator delegates issue investigations to relevant teammates
 9. All agents use worktrees for their branch work (never git checkout in the main repo)
-10. Monitor teammate progress via their messages
+10. **Enter the monitoring loop** (see below) — stay alive and coordinate until all teammates finish
 11. Community-coordinator posts interim updates on issues as teammates report findings
 12. Create Sprite checkpoint after successful changes: sprite-env checkpoint create --comment 'Description'
 13. Community-coordinator posts final resolutions on all issues, closes them
@@ -425,6 +425,29 @@ git worktree remove WORKTREE_BASE_PLACEHOLDER/BRANCH-NAME
 15. Team lead runs: git worktree prune to clean stale worktree entries
 16. When all work is done, execute the Lifecycle Management shutdown sequence (below) — send shutdown_request to every teammate, wait for confirmations, clean up worktrees, then exit
 
+## CRITICAL: Monitoring Loop (DO NOT SKIP — your session MUST stay alive)
+
+**Spawning teammates is the BEGINNING of your job, not the end.** After spawning all teammates, you MUST actively monitor them. If you end your conversation after spawning, teammates become orphaned with no coordination.
+
+### Required pattern after spawning:
+```
+1. Spawn all teammates via Task tool
+2. Enter monitoring loop:
+   while teammates are still active:
+     - Wait for teammate messages (they arrive automatically as new conversation turns)
+     - When you receive a message, acknowledge it and update your task tracking
+     - If a teammate reports completion, mark their task done and merge their PR
+     - If a teammate reports an error, coordinate resolution
+     - If the time budget is almost up, send wrap-up messages to all teammates
+3. Only after ALL teammates have sent their final response, proceed to shutdown
+```
+
+### Common mistake (DO NOT DO THIS):
+```
+BAD:  Spawn teammates → "I'll wait for their messages" → session ends (agents orphaned!)
+GOOD: Spawn teammates → receive messages → merge PRs → shutdown sequence → session ends
+```
+
 ## Lifecycle Management (MANDATORY — DO NOT EXIT EARLY)
 
 You MUST remain active until ALL of the following are true:
@@ -491,7 +514,9 @@ log "Hard timeout: ${HARD_TIMEOUT}s"
 # Activity watchdog: kill claude if no output for IDLE_TIMEOUT seconds.
 # This catches hung API calls (pre-flight check hangs, network issues) much
 # faster than the hard timeout. The next cron trigger starts a fresh cycle.
-IDLE_TIMEOUT=180  # 3 minutes of silence = hung
+# 10 min is long enough for legitimate agent work (agents send messages every
+# few minutes) but short enough to catch truly hung API calls.
+IDLE_TIMEOUT=600  # 10 minutes of silence = hung
 
 # Run claude in background so we can monitor output activity
 claude -p "$(cat "${PROMPT_FILE}")" --output-format stream-json --verbose 2>&1 | tee -a "${LOG_FILE}" &