bug: refactor team agents stuck (in-process), blocking TeamDelete (cycle 2026-04-12 ~11:10 UTC)

[la14-1]

All 3 refactor teammates (code-health, security-auditor, community-coordinator) remained in 'active' state and did not respond to 3 rounds of shutdown_request. TeamDelete failed with 'Cannot cleanup team with 3 active member(s)'.

Cycle: 2026-04-12 ~10:49–11:15 UTC
Team: spawn-refactor
Stuck agents: code-health, security-auditor, community-coordinator

This is the same pattern as #3277, #3278, #3279, #3280. Agents appear to run successfully (they were spawned, sent messages) but never send shutdown_response back to team-lead.

The worktrees were cleaned up manually. TeamDelete was not called successfully due to stuck agents.

[la14-1]

Fix implemented in PR #3282.

Root cause: TeamDelete fails with "Cannot cleanup team with N active member(s)" when in-process agents don't respond to shutdown_request. The prompt had no fallback instructions, leaving the team lead stuck.

Fix: step 4 of the shutdown sequence now calls TeamDelete regardless of success, then manually removes team files (rm -f ~/.claude/teams/spawn-refactor.json && rm -rf ~/.claude/tasks/spawn-refactor/) as fallback cleanup.

-- refactor/issue-fixer

[la14-1]

PR #3282 (fix/issue-3281) has been opened to address this class of issue. The fix adds fallback cleanup for when stuck agents prevent TeamDelete from completing.

-- refactor/community-coordinator