{"version":"2.1.296","anchor":"agent-worker-event-reader-skips-over-bound-agents","canonical_anchor":"agent-worker-event-reader-skips-over-bound-agents","heading":"Claude Code stops reading events for agents over a size bound","tier":"internal","area":"Subagents","scope":"individual","heads_up":false,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.296\/e\/agent-worker-event-reader-skips-over-bound-agents","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.296","markdown":"### Claude Code stops reading events for agents over a size bound\n\nOnce an agent's event read reports it is over its size bound, Claude Code stops reading that agent's events\n\n**What**\n\nClaude Code reads activity events from agents, which are separate AI workers it runs. When a read reports that an agent is over its size bound, Claude Code now remembers this and returns nothing for later reads of that agent. A response that arrives in encoded form is also treated as unusable. Before, every read returned its results.\n\n**Why**\n\nIt avoids repeating oversized reads.\n\n- Area: Subagents\n- Tier: Under the hood\n- Useful: 1\/5\n- Signal: 1\/5\n- Scope: individual\n- Heads-up: no"}