v2.1.199
Claude CodeSummary
This release significantly improves reliability for streaming responses and subagent operations, ensuring partial work is preserved during errors and API failures are correctly reported. Several critical bugs related to background agent stability on Linux and macOS have been resolved. Additionally, transient server rate limits are now automatically retried for subscribers, enhancing workflow continuity.
New Features
- Enabled stacked slash-skill invocations (up to 5) to load all leading skills instead of just the first.
Bug Fixes
- Fixed SSL certificate errors (TLS-inspecting proxies, missing NODE_EXTRA_CA_CERTS, expired certs) causing unnecessary retries; they now fail immediately with a fix hint.
- Fixed streaming responses being discarded when the API emits a mid-stream overloaded/server error; the partial output is now kept with an incomplete-response notice.
- Fixed subagents silently failing instead of returning partial work to the parent when cut off by a rate limit or server error.
- Fixed subagents incorrectly reporting API errors (like usage limit reached) as successful results; errors are now reported to the parent agent.
- Fixed the background-agent daemon on Linux killing itself and all running agents due to a corrupted worker record after an unclean shutdown.
- Fixed background agents failing to cold-start over SSH on macOS with "Could not switch to audit session" (regression in 2.1.196).
- Fixed 'claude stop' being silently undone when it raced a background-agent respawn; respawns now honor the stop command.
- Fixed background job progress indicators stalling for minutes while the job executed long commands.
- Fixed background sessions on memory-starved machines showing a generic error; they now indicate low memory and suggest freeing resources.
- Fixed remote sessions briefly flapping between Working and Idle in the agent view when a background agent completed.
- Fixed idle subagents vanishing from the agent panel while other subagents were still working; surplus idle agents now collapse into an expandable summary row.
- Fixed typing '/model' or '/fast' while viewing a subagent silently opening the lead's model picker; a notice now explains the command applies to the lead.
- Fixed 'SessionStart', 'Setup', and 'SubagentStart' hooks silently hiding stderr when exiting with code 2; the error is now shown in the transcript.
- Fixed 'claude --dangerously-skip-permissions daemon <subcommand>' being treated as a chat prompt instead of running the subcommand.
- Fixed 'SendMessage' silently misrouting when a re-spawned agent reused a previous agent's name; the tool now detects the mismatch and prompts the caller to retarget.
- Fixed opening or resuming a session with no new messages needlessly growing the transcript file.
- Fixed backgrounding a session with '←' or '/background' dropping its '/color' setting from the agent view row.
- Fixed resetting a corrupted config file from the startup recovery dialog destroying it unrecoverably; the file is now backed up first.
- Fixed Claude in Chrome repeatedly opening the reconnect page when sessions run from different builds or config directories.
- Fixed plan mode not prompting for state-changing browser tool calls; read-only 'browser_batch' calls are now correctly auto-allowed.
Improvements
- Transient server rate-limit errors (429s unrelated to your usage limit) are now automatically retried with backoff for subscribers instead of failing the turn.
- Increased the default retry count for non-capacity transient errors via CLAUDE_CODE_RETRY_WATCHDOG to 300 and lifted the cap of 15 on CLAUDE_CODE_MAX_RETRIES.
- 'claude agents' session rows now show pull-request links as bare '#N' without the redundant "PR" label.