RepoJournal
Anthropic @anthropics
123 wires and counting

$ follow Anthropic

Keep up with Anthropic in about 3 minutes: what actually shipped — the commits, pull requests, releases, and security advisories that matter.

or

fair warning: these emails are deeply technical. diffs, version numbers, CVEs, benchmark deltas. if that's not your idea of a good read, this isn't your newsletter.

Folds into your digest — weekly by default, monthly if you prefer. Unsubscribe in one click.

$ status

wire 2026-07-18
stories 10
GitHub RSS Archive Reviews Topics

© 2026 RepoJournal Home Showcase Explore How it works Privacy

$ the-wire · showcase

CLAUDE CODE CLOSES PERMISSION BYPASS; AGENT SDK TIGHTENS VALIDATION

By RepoJournal · Filed · About Anthropic · Composed from the cited sources · methodology

ashwin-ant github-actions[bot] GitHub Actions
3 people shipped this
ashwin-ant @ashwin-ant 2 cited github-actions[bot] @github-actions[bot] 1 cited GitHub Actions @actions-user 1 cited

Claude Code 2.1.214 ships critical security fixes for permission checks across Windows and Bash, while the TypeScript agent SDK now rejects invalid permission modes instead of silently accepting them.

The Claude Code team patched three distinct permission-check vulnerabilities in v2.1.214 [1]. A Windows PowerShell 5.1 permission bypass is now closed, Bash permission checks no longer mishandle file-descriptor redirects, and overly-long commands that broke the permission analyzer are now handled correctly. Separately, a logic error in single-segment `dir/**` allow rules was fixed, preventing unintended auto-approvals for nested directories outside the target scope [1]. On the SDK side, the TypeScript agent SDK v0.3.214 [2] now rejects unrecognized permission modes with an error instead of silently adopting them, hardening the validation layer. The agent SDK also added optional `subkind: 'scheduled-trigger'` support to mark task notifications fired from user-configured scheduled tasks [2]. Both the Python agent SDK [4] and Claude Code Action bumped to these latest versions, ensuring consistency across the toolchain. This coordinated release cycle (Claude Code 2.1.214, Agent SDK TypeScript 0.3.214, Python SDK 0.2.122) reflects a tightening of permission semantics across the platform [3].

Action items

References

  1. [1] v2.1.214 ↗ anthropics/claude-code
  2. [2] v0.3.214 ↗ anthropics/claude-agent-sdk-typescript
  3. [3] chore: bump Claude Code to 2.1.214 and Agent SDK to 0.3.214 ↗ anthropics/claude-code-action
  4. [4] v0.2.122 ↗ anthropics/claude-agent-sdk-python

$ ls anthropics/reviews/ # the step-back reads

the week in review Claude Code 2.1.277 adds AGENTS.md support the month in review Claude Code sync stabilizes, Anthropic SDKs hit 1.0, plugins roll at scale
Topics: AI / ML Full archive → Weekly reviews →

Quick answers

What shipped in Anthropic on July 18, 2026?
Claude Code 2.1.214 ships critical security fixes for permission checks across Windows and Bash, while the TypeScript agent SDK now rejects invalid permission modes instead of silently accepting them. In total, 6 commits and 4 releases landed.
Who contributed to Anthropic on July 18, 2026?
3 developers shipped this update, including github-actions[bot], GitHub Actions, and ashwin-ant.
What were the notable Anthropic updates?
v2.1.214, v0.3.214, and chore: bump Claude Code to 2.1.214 and Agent SDK to 0.3.214.
♥

$ log @anthropics --recent

Verbatim prompt mode closes an @-path file read in the Python SDK 2 shipped Claude Code 2.1.277 reads AGENTS.md, agent-sdk ships the AGENTS.md mod 23 shipped Sandbox-runtime closes a silent deny-glob hole and two Linux startup failures 80 shipped Sales plugin declares Salesforce and Microsoft 365 connectors, plugin directory adds six integrations 67 shipped

$ log --all # elsewhere on the wire

@openclaw Cold-start Gateway diagnostics stop crying wolf, doctor repairs plugin drift Agentic Coding Cline desktop bundles shrink, goose unblocks its own clicks

$ ls showcases/

Rails Django Laravel Node.js Go Vercel Vue.js Kubernetes HashiCorp Arch Linux Linux OpenAI

$ init your-desk

RepoJournal can write one of these for any repo you care about. Sign in with GitHub, pick your repos and sources, and your desk gets a wire like this one every morning.