The Wire · Showcase
CLAUDE CODE 2.1.215 SHIPS WITH SKILL BEHAVIOR CHANGES
By RepoJournal · Filed · About Anthropic
Claude Code no longer auto-invokes verify and code-review skills - you now call them explicitly when needed.
The Claude Code v2.1.215 release [1] marks a shift in how the agent handles code validation. Previously, Claude would run /verify and /code-review automatically; now these are opt-in via explicit commands. This change propagates across the entire agent ecosystem. The TypeScript SDK pushed v0.3.215 [2] to maintain parity with Claude Code, while the Python SDK [3] bundled the updated CLI at v2.1.215. Both the code-action [4] and code-base-action [5] repositories synced the changes immediately. If you're running Claude Code in CI/CD pipelines or automated workflows, this is a breaking change that requires explicit skill invocation.
Action items
- → Update Claude Code to v2.1.215 and review any automation that relied on auto-verify anthropics/claude-code [plan]
- → Sync TypeScript SDK to v0.3.215 if you use claude-agent-sdk-typescript anthropics/claude-agent-sdk-typescript [plan]
- → Audit automated code review workflows - verify and code-review now require explicit invocation anthropics/claude-code [monitor]
References
- [1] v2.1.215 ↗ anthropics/claude-code
- [2] v0.3.215 ↗ anthropics/claude-agent-sdk-typescript
- [3] v0.2.123 ↗ anthropics/claude-agent-sdk-python
- [4] v1.0.178 ↗ anthropics/claude-code-action
- [5] Sync from claude-code-action base-action@af0559e anthropics/claude-code-base-action