{"version":"2.1.284","anchor":"subagent-results-now-include-the-full-hand-back-report","canonical_anchor":"subagent-results-now-include-the-full-hand-back-report","heading":"Subagent results carry the full hand-back report in a new handbackReport field","tier":"internal","area":"Agents","url":"https:\/\/changelogs.core-directive.com\/v\/2.1.284\/e\/subagent-results-now-include-the-full-hand-back-report","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.284","markdown":"### Subagent results carry the full hand-back report in a new handbackReport field\n\nAgent tool results and hand-back recipients now receive the subagent's untruncated report, with any auto-mode security warning\n\n**Unclear.** It is not clear whether the SubagentHandback tool is available to everyone or switched on only for some accounts.\n\n**What**\n\nA subagent is a helper agent Claude starts for a piece of work. It reports back through the `SubagentHandback` tool, and the result already said whether that hand-back was sent, flagged or withheld (`handback`).\n\n- The Agent result now also carries `handbackReport`, taken from the last hand-back's report. It is meant for a client to show instead of `content` when a hand-back was sent or flagged.\n\n- Its `text` is the subagent's whole report, never truncated. Anything in it that imitates Claude Code's own markup has backslashes inserted.\n\n- Its optional `warning` holds auto mode's SECURITY WARNING, or a note that the review could not run.\n\n- When a hand-back recipient is worked out, the report now travels to that recipient along with the target.\n\n- The description of `handback` was rewritten.\n\n**Why**\n\nApps built on Claude Code can show a subagent's full report as its own item, together with any security warning, instead of only as ordinary tool output.\n\n- Area: Agents\n- Tier: Under the hood\n- Useful: 3\/5\n- Signal: 3\/5"}