Claude Code 2.1.269 ships plugin evaluation runner and workflow concurrency controls

On September 12, 2026, Anthropic released Claude Code 2.1.269. The headline feature is a new claude plugin eval command: it runs a plugin's evaluation suite against Claude Code and emits scored, reproducible results as both JSON and HTML reports. That turns plugin quality from a vibe check into something a CI job can gate on, which matters now that organizations ship plugins through managed settings. The same release opens up workflow concurrency: CLAUDE_CODE_WORKFLOW_MAX_CONCURRENT_AGENTS, with a range of 1 to 256, raises the per-run agent limit for the Workflow tool in inference-bound fan-out scenarios. Users can list and switch output styles via the /output-style command, including over Remote Control and in headless and cloud sessions. The release also adds OTEL_METRICS_INCLUDE_REPOSITORY, an environment variable that tags OpenTelemetry metrics and events with vcs.* repository attributes for cleaner attribution in multi-repo environments. Two smaller additions: the Bash tool result now includes a diff of the files a command changed when it handles file edits (bashEditDiffEnabled), and CLAUDE_CODE_GATEWAY_MODEL_DISCOVERY_TIMEOUT_MS extends the LLM gateway /v1/models discovery timeout past its 3-second default.