Claude Report — 2026-06-02

🚀 Anthropic Official

Claude Code v2.1.158 — Auto Mode on Bedrock, Vertex, and Foundry

Auto mode is now available on Amazon Bedrock, Google Vertex AI, and Microsoft Foundry for Opus 4.7 and Opus 4.8; enable with the environment variable CLAUDE_CODE_ENABLE_AUTO_MODE=1.

Claude Code v2.1.160 — Security Guardrails and Ultracode Rename

Introduces permission prompts before writing to shell startup files (.zshenv, .zlogin, .bash_login, ~/.config/git/) and before acceptEdits overwrites build-tool config files (.npmrc, .yarnrc*, bunfig.toml, .bazelrc, .pre-commit-config.yaml, .devcontainer/); renames the dynamic-workflow activation keyword from workflow to ultracode; single-file grep/egrep/fgrep commands now satisfy the read-before-edit requirement; fixes WSL clipboard interop via PowerShell, background-session history loss on restore, voice mode with non-ASCII paths, CJK IME positioning, and auto-mode unavailability messaging on Bedrock/Vertex/Foundry.

Expanding Project Glasswing — 150 New Orgs, $100M Mythos Credits

Adds ~150 organizations across 15+ countries targeting underrepresented critical-infrastructure sectors: energy, water utilities, healthcare, communications, and hardware; the initial 50 partners surfaced 10,000+ high/critical-severity flaws using Claude Mythos Preview since April, and Anthropic commits up to $100M in Mythos usage credits plus $4M in direct donations to open-source security organizations.

Anthropic Confidentially Files S-1 with the SEC

Anthropic submitted a draft Form S-1 on June 1, 2026 for a proposed IPO targeting as early as October 2026; revenue run-rate reached $47B annualized as of May 2026, with a $965B post-money valuation from the recent $65B Series H, putting Anthropic ahead of OpenAI's $852B March 2026 valuation.

🔌 Claude Code Plugins

Plugin tracking skipped this execution (top-10 install cache not loaded).

🛠️ Skills

Piebald-AI/claude-code-system-prompts — Synced to v2.1.160

Updates to Claude Code v2.1.160 (June 2, 2026), tracking all built-in tool descriptions and the updated sub-agent prompt structure introduced in recent releases; the repo self-syncs within minutes of each Claude Code release, adding over 10,500 tokens of new prompt content in this pass.

Unchanged in window: anthropics/knowledge-work-plugins (last commit May 29), VILA-Lab/Dive-into-Claude-Code (last updated May 29), ComposioHQ/awesome-claude-skills (no window activity verified), anthropics/claude-cookbooks (no window activity verified)

🤖 Agents & Subagents

claude-agent-sdk-typescript v0.3.160 — Hook Abort-Signal Fix

Fixes SDK hook callbacks swallowing abort signals: aborting during a PostToolUse hook now ends the turn cleanly with a final result message instead of hanging the calling process (June 2, 2026, in parity with Claude Code v2.1.160).

claude-agent-sdk-typescript v0.3.158 — Bedrock/Vertex/Foundry Auto Mode Parity

Bumps in parity with Claude Code v2.1.158 (May 30, 2026), enabling auto mode support on Amazon Bedrock, Google Vertex AI, and Microsoft Foundry for Opus 4.7/4.8 SDK consumers when CLAUDE_CODE_ENABLE_AUTO_MODE=1 is set.

Unchanged in window: claude-agent-sdk-python (no releases in window), nicobailon/pi-subagents (no user-facing changes verified in window), Chachamaru127/claude-code-harness (last active May 29)

🔗 MCPs & Integrations

modelcontextprotocol/servers — Filesystem Dollar-Sign Preservation

Commit 00ce9cf (PR #4225, May 30, 2026) fixes the filesystem server to preserve literal dollar signs in path templates, preventing incorrect shell-variable expansion when $-prefixed strings are passed through the MCP filesystem tool.

modelcontextprotocol/servers — Memory Server Tool Annotations

Commit 64b1cb0 (PR #3874, May 30, 2026) adds tool annotations to all knowledge graph tools in the memory server, improving capability declarations and discoverability for agents using the knowledge graph entities and relations tools.

Unchanged in window: github-mcp-server (last release May 29), awslabs/mcp (last release May 29), anthropics/claude-for-legal (no new activity verified)

💡 Community — Workflows & Ideas

InfoQ: Dynamic Workflows in Claude Code — Full Technical Breakdown

Published June 1, 2026; explains how ultracode-triggered Dynamic Workflows coordinate hundreds of parallel agents with mid-run checkpointing and resumption, covers practical token-cost warnings (multiple HN reports of 62+ concurrent Opus 4.8 agents hitting the 5-hour cap in under 20 minutes), and clarifies availability across Max/Team/Enterprise plans plus Claude API, Bedrock, Vertex, and Foundry.

luongnv89/claude-howto — Updated to v2.1.160

Visual, example-driven guide to Claude Code updated June 2, 2026 to v2.1.160 compatibility; includes 10 tutorial modules with copy-paste config templates, Mermaid diagrams, and a progressive learning path from basic terminal commands to multi-agent workflows.

📰 Quick Mentions

Relevant links that don't warrant a section of their own:

• TechCrunch: Anthropic files to go public — confidential S-1 details and competitive context versus OpenAI and SpaceX
• CNBC: Anthropic confidentially files IPO prospectus with SEC — valuation and revenue run-rate breakdown
• CNBC: Anthropic expands Mythos to 150 additional organizations — partner reaction quotes on Glasswing expansion
• SecurityWeek: Anthropic Expanding Mythos Access to 150 New Organizations — security-community perspective on the Glasswing cohort
• Washington Post: Claude maker Anthropic files with the SEC to go public — mainstream S-1 coverage