New strings in Anthropic's Claude Code binary hint at internal worker processes, task mana

EXHIBIT 909 · URGENT

Aug 22, 2026 00:48 UTC binary @anthropic-ai/claude-code-darwin-arm64

New strings in Anthropic's Claude Code binary hint at internal worker processes, task management, and potential feature flags like 'planModeResume' and 'unreleased'.

READY TO POST:

anthropic's claude code binary has new strings about 'worker results and system notifications' – could be internal task management or a new way claude interacts with background processes.

seeing some interesting feature flags like 'planModeResume' and 'unreleased' in the latest claude code update. wondering what anthropic is cooking up internally.

⚠️ FROM COMPILED BINARY — lower confidence, verify before posting. (May include bundled-dependency strings, not just first-party changes.) @anthropic-ai/claude-code-darwin-arm64 2.1.238 -> 2.1.239: +8649 strings, -9184 strings, 10 interesting [feature_flag] (8) + ${e?NjS:"Every message you send is to the user."} Worker results and system notifications are internal signals, not conversation partners \u2014 never thank or acknowledge them. Summarize new information for the user as it arrives. + <${sF}>${n.length} background ${r} task(s) from the previous session have no completion record. ${l} They have been marked ${e}. ${c} Task ids in this notification beginning with "${_Hs}" are internal scan markers, not tasks.</${sF}> + This artifact action's input arrived carrying internal approval fields that no permission check attached for this call, so nothing was done. If a hook or SDK host is adding internal fields to this tool's input, remove them and retry. + [planModeResume] ignoring unrecognized internal_metadata.worker_permission_mode + client_enabled + guard_enabled + hook_forwarding_not_ready: internal_error + unreleased [endpoint] (2) + /api/claude_code/notification/preferencesW + https://claude.ai/settings/usage?from=cc_cli_limit_message! [sentence] (5) + ${e?NjS:"Every message you send is to the user."} Worker results and system notifications are internal signals, not conversation partners \u2014 never thank or acknowledge them. Summarize new information for the user as it arrives. + <${sF}>${n.length} background ${r} task(s) from the previous session have no completion record. ${l} They have been marked ${e}. ${c} Task ids in this notification beginning with "${_Hs}" are internal scan markers, not tasks.</${sF}> + This artifact action's input arrived carrying internal approval fields that no permission check attached for this call, so nothing was done. If a hook or SDK host is adding internal fields to this tool's input, remove the