CLAUDE·news
PTEN
№ 0552026-06-06

Claude Report — 2026-06-06

🚀 Anthropic Official

Claude Code v2.1.166 — fallbackModel, Glob Deny Rules, Cross-Session Security Hardening

Released June 6, 2026 at 00:55 UTC. Adds fallbackModel managed setting to configure up to three ordered fallback models tried when the primary is overloaded or returns unexpected non-retryable errors (automatic single-turn retry); glob pattern support in deny-rule tool-name position ("*" blocks all tools); relayed cross-session messages stripped of user authority; MAX_THINKING_TOKENS=0, --thinking disabled, and per-model thinking toggle now fully disable thinking on models that think by default; claude update announces target version before downloading. Fixes: recurring "image could not be processed" error and extra token burn, remote sessions stalling after brief backend disruptions, JetBrains 2026.1+ terminal flickering, Shift+non-ASCII characters dropped in Kitty keyboard protocol terminals, PowerShell command validation hangs on Windows, orphaned claude --bg-pty-host spinning at 100% CPU on macOS, plus voice mode, managed settings, background agent session, and terminal rendering regressions.

Anthropic SDKs — Middleware Support Across All Platforms + Claude Opus 4.1 Deprecated

Released June 5, 2026. TypeScript sdk@0.101.0 ships SDK-level middleware support, scopes request timeouts to the inner fetch only (not the full middleware chain), carries stop_details through beta message_delta accumulation, and fixes JSON number parsing with scientific notation; Python v0.106.0 fixes Foundry client copy()/with_options() and preserves $defs when schema root is a $ref. All platform variants — Vertex v0.17.0, Foundry v0.3.0, Bedrock v0.30.0, AWS v0.4.0 — receive the same middleware addition. All variants also mark claude-opus-4-1-20250805 as deprecated; the Anthropic API changelog (Jun 5) confirms retirement on August 5, 2026, with Opus 4.8 as the recommended migration target.

What We Learned Mapping a Year's Worth of AI-Enabled Cyber Threats

Published June 3, 2026. Analysis of 832 banned accounts reveals threat actors now use AI primarily in post-compromise stages (lateral movement, account discovery) rather than initial access; the proportion of medium-to-high-risk actors rose from 33% to 56% year-over-year, reflecting a shift toward more sophisticated AI application; MITRE ATT&CK does not yet adequately capture "agentic orchestration" — autonomous multi-stage attack chains with minimal human intervention — flagging a detection framework gap that defenders must close before the technique proliferates.

🔌 Claude Code Plugins

Plugin cache unavailable — top-10 tracking skipped this run.

🛠️ Skills

None of the top 5 confirmed updated in window.

Unchanged in window: Piebald-AI/claude-code-system-prompts, anthropics/knowledge-work-plugins, VILA-Lab/Dive-into-Claude-Code, ComposioHQ/awesome-claude-skills, anthropics/skills

🤖 Agents & Subagents

claude-agent-sdk-typescript v0.3.166 — Runtime MCP Resource Tools Fix

Released June 6, 2026 at 00:55 UTC. Fixes MCP resource tools being silently dropped for MCP servers added at runtime via the mcp_set_servers control request — agents that dynamically register MCP servers during a session now correctly receive all resource tools from those servers.

Unchanged in window: nicobailon/pi-subagents, Chachamaru127/claude-code-harness

🔗 MCPs & Integrations

None of the top 5 updated in window.

Unchanged in window: modelcontextprotocol/servers, github-mcp-server, awslabs/mcp, modelcontextprotocol/typescript-sdk, portainer/portainer-mcp

💡 Community — Workflows & Ideas

No relevant news today.

📰 Quick Mentions

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

← back to archiveend of edition № 055