{"version":"2.1.91","anchor":"subagent-report-file-prevention-gradual-rollout","canonical_anchor":"subagent-report-file-prevention-gradual-rollout","heading":"Subagent Report File Prevention [Gradual Rollout]","tier":null,"area":null,"scope":null,"heads_up":null,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.91\/e\/subagent-report-file-prevention-gradual-rollout","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.91","markdown":"### Subagent Report File Prevention [Gradual Rollout]\n\n**What**\n\nSubagents are instructed to return findings as text in their final message instead of creating report\/summary `.md` files.\n\n**Status**\n\nFeature-flagged via `tengu_sub_nomdrep_q7k` (default: false)\n\n**Details**\n\n- Adds instruction: \"Do NOT write report\/summary\/findings\/analysis .md files. Return findings directly as your final assistant message \u2014 the parent agent reads your text output, not files you create.\"\n\n**Evidence**\n\nSubagent no-report instruction (gated by `tengu_sub_nomdrep_q7k`, search for `\"report\/summary\/findings\/analysis .md files\"`)\n\n- Flag `tengu_sub_nomdrep_q7k`: On for this account, and not off by default (read for one account on one subscription tier against v2.1.91; this account: on, anonymous baseline: on, compiled default: no gate table built for this version) These values were read against a different version of Claude Code, so treat them as the nearest reading available instead of one taken on this release."}