Claude Code /team-onboarding: Auto-Generate Teammate Ramp-Up Guides
Claude CodeClaude Code v2.1.101 introduces /team-onboarding, a new command that analyzes a developer's local Claude Code usage history and automatically generates a structured ramp-up guide for new teammates. Rather than writing onboarding documentation manually, teams can let Claude Code synthesize its knowledge of how the codebase has been navigated into actionable documentation. The release also improves session resumption, rate-limit messaging, and settings resilience.
Key Takeaways
- /team-onboarding generates onboarding documentation from real usage history, meaning the resulting guide reflects how the codebase is actually navigated rather than how it was documented at a point in time.
- New teammates can ramp up faster because the generated guide captures established workflows, common commands, and project conventions that would otherwise require tribal knowledge transfer.
- --resume now accepts session titles, removing the need to look up opaque session IDs when resuming previous work from the command line.
- Rate-limit error messages are now clearer and more actionable, helping developers distinguish rate limits from other failures and decide how to respond.
- Refusal error messages are more descriptive, reducing silent stalls in agentic workflows where model refusals previously surfaced as cryptic errors.
- Settings persistence reliability improves with a fix for edge cases where configuration changes were not correctly preserved across sessions.
Sources & Mentions
5 external resources covering this update
New Features
/team-onboarding: Instant Ramp-Up Guides for New Teammates
Claude Code v2.1.101 introduces the /team-onboarding command, a new capability that automatically generates structured onboarding documentation for new team members. When a developer runs /team-onboarding, Claude Code analyzes the local usage history — the commands run, files navigated, workflows established — and synthesizes that knowledge into a coherent ramp-up guide that a new teammate can follow.
The practical impact is significant: onboarding documentation is notoriously difficult to keep current, and it is typically written from memory rather than from actual observed behavior. Because /team-onboarding derives its content from real usage data, the resulting guide reflects how the codebase is actually navigated in practice, not how it was documented months ago. Teams adopting this feature can reduce the time new developers spend figuring out project conventions, common workflows, and tool usage patterns.
The command is designed to be run by any developer who has been actively using Claude Code on the project. The more history Claude Code has, the richer the generated guide. The output is a structured markdown document that teams can commit to their repository or share directly.
Improvements
--resume Now Accepts Session Titles
The --resume flag, used to continue a previous Claude Code session from the command line, now accepts session titles in addition to session IDs. This is a meaningful ergonomic improvement: session IDs are opaque strings that require looking up, while session titles are human-readable names set when a session is created. Developers who name their sessions can now resume them without hunting for the correct ID.
Clearer Rate-Limit Messages
When Claude Code encounters API rate limits, it now surfaces clearer, more actionable messages. Previously, rate-limit responses could appear as generic errors that were difficult to distinguish from other failure modes. The improved messages communicate directly that the limit has been reached and give developers better context for deciding whether to wait, switch models, or adjust their usage patterns.
Refusal Error Messages Improved
Similarly, error messages for model refusals have been improved to be more descriptive and less cryptic. When Claude declines to perform an action, the message now more clearly communicates what happened and why, reducing confusion in agentic workflows where refusals can otherwise stall automation pipelines silently.
Settings Resilience Fix
A settings persistence bug has been resolved. In some configurations, user settings were not correctly preserved across sessions — specifically in edge cases involving concurrent writes or corrupted state. This fix improves the reliability of settings persistence, ensuring that configuration changes made in one session carry forward correctly.