archive
Latest editions
June 2026
- № 07930 Jun2026-06-30
Claude Report — 2026-06-30
- Anthropic launches self-hosted Claude apps gateway for Bedrock and Google Cloud — corporate SSO, RBAC, per-user cost attribution, and spend limits for Claude Code teams (Jun 29)
- Claude in Microsoft Foundry is now GA: Opus 4.8 + Haiku 4.5 on Azure, running on NVIDIA GB300 Blackwell Ultra GPUs (Jun 29)
06/08 sections - № 07829 Jun2026-06-29
Claude Report — 2026-06-29
- US government clears Claude Mythos 5 for ~100 companies and federal agencies after a two-week export-control standoff (June 26-27)
- Claude API rate limits raised: Sonnet and Haiku now match Opus at every tier; tiers consolidated to Start / Build / Scale (June 26)
07/08 sections - № 07728 Jun2026-06-28
Claude Report — 2026-06-28
- Epic Games adds official Unreal Engine Skills plugin to Claude Code marketplace (June 26): 30+ toolsets, hundreds of editor tools via MCP
- Anthropic Economic Index "Cadences" (June 26): Claude usage mirrors real-world daily rhythms; heavy delegators most optimistic about their jobs
06/08 sections - № 07627 Jun2026-06-27
Claude Report — 2026-06-27
- US lifts Mythos 5 export block (June 26): 100+ US institutions and agencies gain access; Fable 5 still blocked.
- Anthropic formally accuses Alibaba of record distillation attack: 28.8M Claude exchanges via ~25K fraudulent accounts.
05/08 sections - № 07526 Jun2026-06-26
Claude Report — 2026-06-26
- Anthropic launches Claude Tag for Slack (June 23): async AI teammate with persistent channel memory, beta for Enterprise/Team customers
- Claude Code v2.1.193 (June 25): `autoMode.classifyAllShell`, OpenTelemetry assistant-response logging, live bash path autocomplete, phantom subagent fix
05/08 sections - № 07425 Jun2026-06-25
Claude Report — 2026-06-25
- Anthropic sends Senate letter accusing Alibaba of the largest known Claude distillation attack — 25K fraudulent accounts, 28.8M exchanges used to harvest Qwen capabilities (June 24)
- Claude Code v2.1.191 (June 24): `/rewind` now survives `/clear`; CPU usage drops ~37% during streaming; stopped background agents no longer resurrect
06/08 sections - № 07324 Jun2026-06-24
Claude Report — 2026-06-24
- Claude Tag launches in Slack beta for Enterprise/Team (June 23): multiplayer, context-persistent AI teammate that joins channels and works asynchronously
- Claude Code v2.1.187 (June 23): `sandbox.credentials` blocks credential files and secret env vars from sandboxed commands; org-level model restrictions in picker; CJK mojibake fix
05/08 sections - № 07223 Jun2026-06-23
Claude Report — 2026-06-23
- Claude Code v2.1.186 (June 22): `claude mcp login/logout` CLI commands for MCP auth, Skills tab in /plugin, iTerm2 teammate mode
- claude-agent-sdk-typescript v0.3.186 (June 22): background agents now forward `canUseTool` prompts instead of auto-denying; `rewind_conversation` added
06/08 sections - № 07122 Jun2026-06-22
Claude Report — 2026-06-22
- Nobel laureate John Jumper (AlphaFold, DeepMind) announces move to Anthropic on June 19
- Claude Code v2.1.183 (June 19): auto mode blocks destructive git ops and IaC teardown unless explicitly requested
07/08 sections - № 07021 Jun2026-06-21
Claude Report — 2026-06-21
- Enterprise-Managed MCP Connector Access launches in beta (June 18): Okta-provisioned zero-touch access to Asana, Figma, Linear, and 4 other connectors for Team/Enterprise orgs
- Claude Code Artifacts (June 18): sessions now output live, version-tracked shareable web pages, in beta for Team and Enterprise
07/08 sections - № 06920 Jun2026-06-20
Claude Report — 2026-06-20
- AlphaFold co-creator and 2024 Nobel Chemistry laureate John Jumper announces departure from Google DeepMind for Anthropic (June 19) — biggest AI talent acquisition of the month
- Anthropic opens Seoul office (June 17); NAVER deploys Claude Code across its entire engineering org, Samsung SDS and LG CNS follow with enterprise-wide rollouts
06/08 sections - № 06819 Jun2026-06-19
Claude Report — 2026-06-19
- Claude Code v2.1.183 (June 19) adds auto-mode guardrails that block `git reset --hard`, `terraform destroy`, and similar destructive commands unless explicitly requested
- Artifacts in Claude Code (June 18) turns sessions into live, version-tracked shareable HTML pages — beta for Team and Enterprise plans
07/08 sections - № 06718 Jun2026-06-18
Claude Report — 2026-06-18
- Claude Code v2.1.181 (June 17) adds `/config` prompt syntax, Apple Events support, Bun 1.4, and 30+ bug fixes
- Claude Code v2.1.178 (June 15) ships `Tool(param:value)` permission filtering and nested `.claude/` directory scoping
08/08 sections - № 06617 Jun2026-06-17
Claude Report — 2026-06-17
- Anthropic reverses the Agent SDK billing overhaul on the exact day it was set to take effect (June 15) — subscriptions unchanged until further notice
- `claude-sonnet-4-20250514` and `claude-opus-4-20250514` model IDs retired June 15 — all API calls now return an error; upgrade to Sonnet 4.6 or Opus 4.8
08/08 sections - № 06516 Jun2026-06-16
Claude Report — 2026-06-16
- Claude Code v2.1.178 ships `Tool(param:value)` parameter-based permission rules and nearest-directory `.claude/` resolution — the most impactful permission-system expansion since allow/deny lists launched
- Two new official plugins land on June 15: boltz (biomolecular design via Boltz-2 API) and lovable (Lovable app platform via OAuth 2.1 MCP)
07/08 sections - № 06415 Jun2026-06-15
Claude Report — 2026-06-15
- Anthropic publishes official rebuttal to the Fable 5/Mythos 5 US government suspension, calling the demonstrated bypass "minor and previously known" — not grounds for a full commercial recall
- TCS × Anthropic: Claude deployed to 50,000 TCS employees across 56 countries in banking, insurance, healthcare, and public sector
07/08 sections - № 06314 Jun2026-06-14
Claude Report — 2026-06-14
- DXC deploys Claude across banking, airlines, insurance, and government — OASIS managed-services platform built with 95%+ Claude-generated code
- Anthropic launches Claude Corps: $150M commitment, 1,000 early-career fellows at US nonprofits, $85k/year salary
06/08 sections - № 06213 Jun2026-06-13
Claude Report — 2026-06-13
- US government issues export control directive disabling Fable 5 and Mythos 5 worldwide — first government-forced takedown of a deployed frontier model
- Anthropic acknowledges silent model-switching in Fable 5 for AI/security queries; pledges transparent refusals going forward
07/08 sections - № 06112 Jun2026-06-12
Claude Report — 2026-06-12
- Claude Fable 5 (`claude-fable-5`) goes GA June 9 — Anthropic's most capable widely released model; $10/$50 per M tokens in/out; 1M context; always-on adaptive thinking; available on Claude API, Bedrock, Vertex, Microsoft Foundry, and GitHub Copilot
- Fable 5 brings three integration-breaking changes: `stop_reason: "refusal"` (HTTP 200), ~30% heavier tokenizer vs. pre-Opus 4.7 models, and mandatory 30-day data retention overriding even zero-retention enterprise agreements
07/08 sections - № 06011 Jun2026-06-11
Claude Report — 2026-06-11
- Claude Code v2.1.172 (Jun 10) unlocks nested sub-agents up to 5 levels deep — a two-year hard constraint removed
- Claude Code v2.1.169 (Jun 8) adds `--safe-mode`, `/cd` working-directory migration, and `disableBundledSkills`
07/08 sections - № 05910 Jun2026-06-10
Claude Report — 2026-06-10
- Anthropic launches Claude Fable 5 (June 9): first public Mythos-class model, leads agentic-coding benchmarks, $10/$50 per M tokens, adaptive thinking always-on with no raw thinking output
- Claude Mythos 5 ships simultaneously for Project Glasswing: same weights, safety classifiers lifted for vetted cyber defenders and biology researchers
07/08 sections - № 05809 Jun2026-06-09
Claude Report — 2026-06-09
- Claude Code v2.1.169 (Jun 8) ships `--safe-mode` flag, `/cd` command, and `disableBundledSkills` setting — three new escape hatches for debugging and customization control
- Anthropic launches `ClaudeForFoundationModels` Swift package (Jun 8) at WWDC 2026: Claude as a drop-in `LanguageModel` backend for iOS 27 / macOS 27 apps
07/08 sections - № 05708 Jun2026-06-08
Claude Report — 2026-06-08
- Claude Code v2.1.166 (Jun 6) adds `fallbackModel` for graceful model failover and hardens cross-session messaging against privilege escalation
- Anthropic deprecates Claude Opus 4.1 API (Jun 5) — retirement Aug 5, 2026; migrate to Opus 4.8
06/08 sections - № 05607 Jun2026-06-07
Claude Report — 2026-06-07
- Anthropic publishes first chemistry specialization white paper (Jun 5): Claude Opus 4.7 matches ChemDraw and MestReNova on NMR forward prediction and uniquely enables inverse prediction (spectrum → structure)
- Claude Code v2.1.161-163 (Jun 4) adds `requiredMinimumVersion`/`requiredMaximumVersion` managed settings, `/plugin list`, `agents --json waitingFor`, and hook `additionalContext` support
07/08 sections - № 05506 Jun2026-06-06
Claude Report — 2026-06-06
- Claude Code v2.1.166 (Jun 6) introduces `fallbackModel` configuration (up to 3 cascading fallbacks), glob patterns in deny rules, and cross-session message authority stripping
- All Anthropic SDKs ship middleware support on Jun 5; Claude Opus 4.1 officially deprecated (API retirement Aug 5, 2026 — migrate to Opus 4.8)
05/08 sections - № 05405 Jun2026-06-05
Claude Report — 2026-06-05
- Claude Code v2.1.163 (Jun 4) ships managed version guardrails, `/plugin list` command, and `additionalContext` in `Stop`/`SubagentStop` hooks for non-error Claude feedback
- v2.1.160 (Jun 2, not covered yesterday) renames the dynamic-workflow trigger from `workflow` → `ultracode` and adds safety gates before writing to shell/build-tool config files
05/08 sections - № 05304 Jun2026-06-04
Claude Report — 2026-06-04
- Claude Code v2.1.162 (Jun 3) and v2.1.161 (Jun 2) ship 30+ improvements: `waitingFor` field in `claude agents --json`, Remote Control footer pill, Windsurf → Devin Desktop rename, OTEL metric labels, and per-agent `done/total` progress counters
- Claude API (Jun 2) adds `max_tokens` cap to the advisor tool and eliminates billing for `stop_reason: "refusal"` responses that generate zero output
08/08 sections - № 05203 Jun2026-06-03
Claude Report — 2026-06-03
- Anthropic publishes first systematic MITRE ATT&CK gap analysis for AI-enabled attackers (June 3), covering 832 banned accounts; identifies missing taxonomy for autonomous AI-orchestrated attack chains
- Claude Code v2.1.161 (June 2) fixes a credential leak in `claude mcp`, adds OTEL metric dimension labels, and per-agent `done/total` progress counters
07/08 sections - № 05102 Jun2026-06-02
Claude Report — 2026-06-02
- Claude Code v2.1.158 (May 30) brings auto mode to Bedrock, Vertex, and Foundry for Opus 4.7/4.8
- Claude Code v2.1.160 (June 2) adds security prompts for shell/build-tool config writes; renames dynamic-workflow trigger to `ultracode`
08/08 sections - № 05001 Jun2026-06-01
Claude Report — 2026-06-01
- Claude Code v2.1.157 introduces automatic plugin loading from `.claude/skills` without marketplace
- Auto mode reaches Bedrock, Vertex, and Foundry for Opus 4.7/4.8 in v2.1.158
08/08 sections