Who contributed to Anthropic on May 29, 2026?

7 developers shipped this update, including mhegazy, bryan-anthropic, noahzweben, yamz8, kevintang513, dazzaji, and akhilesharora.

What were the notable Anthropic updates?

Merge pull request #2073 from anthropics/fix-2071-macos-python-39, security-guidance: enable LLM review on default macOS Python 3.9 (#2071), and feat(scan): emit per-entry sticky verdict comments.

@anthropics

Claude SDKs and developer tooling

github ↗

Pick a date

Topics: AI / ML Full archive →

The Wire · Showcase

MACOS PYTHON 3.9 BUG SILENCES SECURITY REVIEWER - AGENT SDK HOSTING COOKBOOK SHIPS

By RepoJournal · Filed 06:02 UTC on May 29, 2026 · About Anthropic

7 people shipped this

mhegazy @mhegazy 1 cited

bryan-anthropic @bryan-anthropic 1 cited

noahzweben @noahzweben 1 cited

yamz8 @yamz8 1 cited

kevintang513 @kevintang513 1 cited

dazzaji @dazzaji 1 cited

akhilesharora @akhilesharora 1 cited

The security-guidance plugin's agentic commit reviewer has been silently failing on macOS systems with Apple's default Python 3.9, even when Python 3.10+ is installed.

A three-layer bug in claude-plugins-official [2] caused the LLM-powered commit reviewer to never run on affected macOS machines. The issue: `sg-python.sh` only checked major version numbers, always preferring Apple's 3.9 over newer Homebrew installs, while `claude_agent_sdk` requires Python 3.10+. The fix merged overnight [1] and patches the version detection logic. In parallel, the same repo shipped per-entry verdict comments for plugin scans [3], allowing downstream dashboards and label automation to read verdicts directly from PRs instead of scraping logs. The team also added Apache 2.0 licensing at the repo root for consistency [4]. Over in claude-cookbooks, the Agent SDK hosting cookbook just dropped [6], walking teams through deployment across Docker Compose, Modal, and Kubernetes using identical container images and HTTP interfaces. On the legal tools side, two critical fixes landed: a handoff request extractor that was truncating nested JSON payloads at the first closing brace [7], and a Linux compatibility fix for the deployment script where mktemp templates needed three X's instead of two [8]. The healthcare desk documented the ICD-10 Codes MCP connector in the main README [5], closing a gap where users couldn't discover it from top-level docs.

Action items

→ If you use claude-plugins-official on macOS, pull the Python 3.9 fix immediately anthropics/claude-plugins-official [immediate]
→ Review the new Agent SDK hosting cookbook for deployment patterns before your next rollout anthropics/claude-cookbooks [plan]
→ If running claude-for-legal on Linux, upgrade to the mktemp fix anthropics/claude-for-legal [immediate]

References

[1] Merge pull request #2073 from anthropics/fix-2071-macos-python-39 anthropics/claude-plugins-official
[2] security-guidance: enable LLM review on default macOS Python 3.9 (#2071) ↗ anthropics/claude-plugins-official
[3] feat(scan): emit per-entry sticky verdict comments ↗ anthropics/claude-plugins-official
[4] Add Apache 2.0 LICENSE to repo root ↗ anthropics/claude-plugins-official
[5] Add ICD-10 Codes MCP to README documentation ↗ anthropics/healthcare
[6] Agent SDK cookbook 07: Hosting your agent (Docker / Modal / Kubernetes) ↗ anthropics/claude-cookbooks
[7] Fix handoff_request extractor truncating nested payloads at the first `}` ↗ anthropics/claude-for-legal
[8] Fix mktemp template so deploy-managed-agent.sh runs on Linux ↗ anthropics/claude-for-legal

Quick answers

What shipped in Anthropic on May 29, 2026?: The security-guidance plugin's agentic commit reviewer has been silently failing on macOS systems with Apple's default Python 3.9, even when Python 3.10+ is installed. In total, 11 commits and 10 pull requests landed.
Who contributed to Anthropic on May 29, 2026?: 7 developers shipped this update, including mhegazy, bryan-anthropic, noahzweben, yamz8, kevintang513, dazzaji, and akhilesharora.
What were the notable Anthropic updates?: Merge pull request #2073 from anthropics/fix-2071-macos-python-39, security-guidance: enable LLM review on default macOS Python 3.9 (#2071), and feat(scan): emit per-entry sticky verdict comments.

CLAUDE AGENT SDK FIXES STDIN CLOSURE BUG IN BACKGROUND TASKS

The Python SDK's query method was prematurely closing stdin when background subagents were still in flight, breaking tool calls and silently bypassing hooks - now fixed in v0.2.127.

ai 16 shipped 1-min read

@anthropics 1 day ago

BUFFA v0.9.1 PATCHES REGRESSION THAT BROKE ORDINARY SCHEMAS

Buffa shipped an emergency patch overnight after v0.9.0 broke codegen on schemas with a few hundred proto files, small enough to hit in production.

ai 113 shipped 1-min read

@anthropics 2 days ago

CLAUDE FOUNDATION MODELS ADDS APP ATTEST, PLUGIN ECOSYSTEM STABILIZES

iOS developers can now ship Claude calls without API keys in the binary, backed by Apple's Secure Enclave.

ai 111 shipped 1-min read

@anthropics 3 days ago

BUFFA FIXES V0.9.0 REGRESSION, CLAUDE PLUGINS ECOSYSTEM GAINS 9 NEW INTEGRATIONS

Buffa's codegen was silently rejecting moderately large schemas due to a memory bound meant only for untrusted input, but a fix shipped overnight restores throughput and clears the path for googleapis-scale proto sets.

ai 156 shipped 2-min read

Elsewhere on the wire

AI Agents about 9 hours ago

CLAUDE OPUS 5 LANDS ACROSS THE STACK

The newest Anthropic model is now live in langchain, Cline, and llama-index, with native support for extended reasoning and 1M context windows.

ai-agents 28 shipped 1-min read

Local LLMs about 9 hours ago

OLLAMA LANDS LAGUNA SUPPORT AND CRUSHES MEMORY LEAKS WHILE SGLANG HITS V0.5.16 WITH CONFIDENCE-DRIVEN SPECULATIVE DECODING

Ollama shipped three critical performance and reliability fixes for Metal residency and concurrent access patterns, while SGL-Lang released 0.5.16 with a new speculative algorithm hitting 383.7 tok/s on DeepSeek-V4.