Sign in
← AI PulseSep 11, 2026

Deep · news · Single-source brief

Claude Code Updates Include Plugin Evaluation and Concurrency Controls

Recent updates to Claude Code introduce a plugin evaluation suite, new output style controls, and enhanced OpenTelemetry integration.

By Illumora Editorial

Source · Sep 11, 2026, 7:17 PM · On Illumora · Sep 11, 2026, 7:22 PM

Media from the primary source — shown here so you can stay on Illumora.

Rewritten from one allowlisted primary — not independent enterprise reporting. Lanes →

Brief drafted by Illumora’s editorial model from the linked primary source. Ops desk reviews flagged pieces. How we write →

Read the source →Claude Code What's New + releases — Claude Code changelog - Claude Code Docs
Save

Recent updates to Claude Code introduce a plugin evaluation suite, allowing users to run a plugin's evaluation suite against Claude Code to obtain scored, reproducible results. These results are provided in both JSON and HTML report formats.

Key Points

  • A claude plugin eval command has been added to run a plugin's evaluation suite against Claude Code, generating scored, reproducible results in JSON and HTML formats.
  • The /output-style [name] command now lists and switches output styles, including via Remote Control and in cloud or headless sessions.
  • A diff of file changes made by a Bash command is now included in the Bash tool result when the bashEditDiffEnabled setting is active.
  • OTEL_METRICS_INCLUDE_REPOSITORY tags OpenTelemetry metrics and events with vcs. repository attributes, with commit events receiving vcs.ref.head. when OTEL_LOG_TOOL_DETAILS is enabled.
  • CLAUDE_CODE_GATEWAY_MODEL_DISCOVERY_TIMEOUT_MS extends the LLM gateway /v1/models discovery timeout, which defaults to 3 seconds.
  • A spinner tip suggests using /focus for a view displaying only the prompt, a one-line work summary, and the response.
  • CLAUDE_CODE_WORKFLOW_MAX_CONCURRENT_AGENTS (ranging from 1 to 256) raises the Workflow tool's per-run concurrent agent limit for inference-bound fan-outs.
  • A fix addresses partial invalidation of the prompt cache on the turn following a response cut off at the output-token limit and automatically resumed.

Context

According to the Claude Code changelog, these additions aim to enhance the development and debugging experience for users. The plugin evaluation suite provides a structured method for assessing plugin performance and reliability. The new output style controls offer greater flexibility in how information is presented across different environments. OpenTelemetry integration improvements allow for more detailed monitoring and observability of repository-related events.

Why It Matters

These updates provide builders with more robust tools for evaluating, customizing, and monitoring their Claude Code workflows. The ability to evaluate plugins directly and control concurrent agent limits can streamline development and deployment, while enhanced observability features support better debugging and performance analysis.

What To Do

  • Review the claude plugin eval --help documentation to understand the new plugin evaluation capabilities.
  • Experiment with the /output-style [name] command to explore different output presentations in various Claude Code sessions.
  • Note the CLAUDE_CODE_WORKFLOW_MAX_CONCURRENT_AGENTS setting for optimizing inference-bound fan-out scenarios in the Workflow tool.
  • Examine the OpenTelemetry integration details, particularly OTEL_METRICS_INCLUDE_REPOSITORY and OTEL_LOG_TOOL_DETAILS, for enhanced monitoring of code repositories.