{"version":"2.1.307","anchor":"the-agent-tool-description-in-the-reminder-read-offset-promp","canonical_anchor":"the-agent-tool-description-in-the-reminder-read-offset-promp","heading":"Agent tool guidance rewritten in the file-offset reminder prompt","tier":"notice","area":"Subagents","scope":"individual","heads_up":false,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.307\/e\/the-agent-tool-description-in-the-reminder-read-offset-promp","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.307","markdown":"### Agent tool guidance rewritten in the file-offset reminder prompt\n\nIn the file-offset reminder prompt, Claude's helper-agent instructions now say to delegate reviews only for an independent view, giving code, not conclusions\n\n**What**\n\nThe description of the `Agent` tool was rewritten in the prompt Claude Code sends with its reminder about reading a file from a given starting point. The `Agent` tool is how Claude passes work to a subagent, a separate helper copy of Claude.\n\nThe new text includes guidance on review work: Claude should delegate a review only when it wants an opinion that isn't anchored on its own, and then give the helper the code rather than its own conclusion.\n\n**Why**\n\nThis text is part of Claude's own instructions, so the rewrite changes how Claude uses helpers for reviews. A helper given only the code can reach its own view instead of repeating Claude's.\n\n- Area: Subagents\n- Tier: You'll notice\n- Useful: 1\/5\n- Signal: 1\/5\n- Scope: individual\n- Heads-up: no"}