AI Agent News/stories/claude-code-v2-1-312-adds-an-oidc-user-id-claim-and-fixes-session-and-tool-bugs.md

Frameworks

Claude Code v2.1.312 adds an OIDC user ID claim and fixes session and tool bugs

Anthropic has released Claude Code v2.1.312, a point release that adds an OIDC user ID claim option to the Claude apps gateway and fixes a range of session, tool and cloud bugs. The changelog also changes how hooks are limited, how some custom model names are handled, and how usage limit messages are worded.

Source: github.com/anthropics/claude-code/releases/tag/v2.1.312

The release adds oidc.user_id_claim to the Claude apps gateway, which lets it identify people by an id_token claim other than sub, such as Microsoft Entra ID's oid. Fixes include managed settings and organization policy being fetched again after signing in from the VS Code extension, and skills being found after a Bash cd into a deeper directory of a git repository.

Other fixes cover the Read tool failing on some PDF pages with a pdftoppm error, a command with an invalid name stopping other commands from loading, and Code Review reviewing an outdated commit when GitHub delivered a push event late. Prompt and agent hooks now run at most 24 at a time per session, including sub-agents, and Bedrock, Vertex, Foundry and Mantle use a shorter system prompt for unrecognized model names unless CLAUDE_CODE_SIMPLE_SYSTEM_PROMPT=0 is set.

Why it matters if you build agents

Teams whose agent setups start many prompt or agent hooks at once should check the new limit of 24 concurrent hooks per session. Anyone using custom deployment names or inference profile ARNs on Bedrock, Vertex, Foundry or Mantle will get the shorter system prompt by default, so they should set CLAUDE_CODE_SIMPLE_SYSTEM_PROMPT=0 if they need the full one.

Get stories like this every weekday

One short email with the day's agent tooling news.

Subscribe