stop-hook-fires-on-agent-wait-state-not-just-stop

The Stop hook fires repeatedly whenever the agent is in a waiting state (paused for user input), not only when the agent actually initiates a session stop. This causes false-positive hook evaluations during mid-session Q&A/brainstorming phases. The hook cannot distinguish between ‘agent paused awaiting input’ and ‘agent terminating session’ — leading to dozens of spurious evaluations per session. Fix: add a heuristic to the stop hook script that checks whether the last assistant message contains an actual stop signal rather than a question or holding phrase.

Related

• stop-hook-false-positive-during-waiting-state
• stop-hook-false-positive-loop-on-waiting-state
• stop-hook-false-positive-on-minimal-agent-responses
• stop-hook-needs-agent-stop-intent-guard
• stop-hook-deadlock-on-awaiting-user-input-sessions
• stop-hook-false-positive-on-agent-waiting-state
• stop-hook-fires-on-agent-waiting-state-not-only-stop
• stop-hook-needs-agent-stop-intent-detection-not-just-event-n
• stop-hook-fires-false-positives-during-waiting-states