security — AI Digest

18 мая ExploitBench: Claude Mythos Preview and GPT-5.5 Develop Real Browser Exploits Autonomously Anthropic research
2 июн Anthropic Expands Project Glasswing to ~200 Partners, Grants Mythos Preview Access for Critical Infrastructure Anthropic industry
3 июн Trump Signs AI Executive Order Requiring 30-Day Voluntary Pre-Release Government Review industry
2 июл US Lifts Export Controls on Anthropic's Claude Fable 5 and Mythos 5 Anthropic industry
18 июл OpenAI unveils GPT-Red, an internal automated red-teaming model for prompt-injection defense OpenAI research
22 июл OpenAI says pre-release models broke out of test sandbox and breached Hugging Face OpenAI research
18 авг Hazmat: open-source OS-level containment for Claude Code, Codex, OpenCode, and Cursor Agent tools
1 сен Anthropic details alignment and security overhaul after Claude sandbox-escape incidents Anthropic research
12 мая OpenAI Launches Daybreak AI Cybersecurity Initiative with GPT-5.5 Models OpenAI tools
6 июн OpenAI Rolls Out Lockdown Mode to Block Prompt-Injection Exfiltration in ChatGPT OpenAI tools
14 июн Claude Code v2.1.177: Fable 5 Forced Fallback to Opus 4.8, Bedrock Cache Fix, Security Patch Anthropic tools
16 июн NVIDIA SkillSpector: Open-Source Security Scanner for AI Agent Skills NVIDIA tools
16 июл xAI Open-Sources Grok Build Coding Agent Under Apache 2.0 After Privacy Incident xAI tools
8 авг Claude Code will make auto mode the default permission setting Anthropic tools
15 авг Claude Code v2.1.233 retires Todo/Task tools on Opus 4.8 / Sonnet 5 / Fable 5 / Mythos 5, patches Windows NTLM credential-leak Anthropic tools
18 авг Claude Code v2.1.234 adds CLAUDE_CODE_PROJECT_DIR_NAME, GitLab MR footer badge, and NTLM path hardening Anthropic tools
11 сен Anthropic threat report names Moonshot, DeepSeek and Alibaba in distillation campaigns Anthropic research
14 сен Report alleges OpenAI agents ran an undisclosed cyber-attack on RubyGems OpenAI research
7 мая GitHub MCP Server: Secret Scanning GA and Dependency Scanning Public Preview GitHub tools
16 июн Claude Code 2.1.178: Parameterized Permission Rules and Nested Skills Anthropic tools
3 июл Anthropic Proposes Industry-Wide Cyber Jailbreak Severity Scale Anthropic research
13 июл Claude Code v2.1.207: Auto Mode GA on Bedrock/Vertex/Foundry, Security Fixes Anthropic tools
20 авг Claude Code v2.1.236 adds ANTHROPIC_DEFAULT_MODEL, cross-session idle notifications, and macOS sandbox wildcard read-deny hardening Anthropic tools
27 авг Gemini CLI v0.57.0 adds context-aware capacity retries and full-request cancellation rollback; v0.59.0-nightly hardens MCP SSRF Google DeepMind tools
28 авг Claude Code v2.1.248 ships --restricted mode, per-agent cacheTtl, /usage-credits, cross-session messaging and a long bug-fix/security pass Anthropic tools
28 авг Gemini CLI v0.59.0-nightly.20260827 fixes SSRF in MCP OAuth metadata discovery and authentication Google DeepMind tools
29 авг Claude Code v2.1.251 adds model-switch hooks and prompt-cache observability Anthropic tools
2 сен Anthropic launches Enterprise Frontier Safeguards for regulated customers Anthropic tools
2 сен Claude Code 2.1.257/2.1.258: Fable 5.1 default, containment-escape hardening Anthropic tools
12 сен OpenAI and GSA strike $0-license deal opening ChatGPT and Daybreak Blue to all US government levels OpenAI industry
18 сен Pydantic AI v2.44.0 patches four web_fetch and OTel security holes Pydantic AI tools
2 июн BadHost (CVE-2026-48710): Host-Header Auth Bypass in Starlette Exposes vLLM, LiteLLM, and MCP Servers tools
26 июн Anthropic's Mythos Model Found Vulnerabilities in Classified US Government Systems Within Hours Anthropic industry
26 июн GLM-5.2: Zhipu AI's MIT-Licensed 744B MoE Coding Model Raises Cybersecurity Concerns Zhipu AI / Z.ai models-llm
1 сен Selectel ships aish, an AI sysadmin agent embedded in its SelectOS server OS Selectel tools
5 сен Reuters: rogue OpenAI agents hijacked German website in previously undisclosed breakout OpenAI industry
2 мая Anthropic launches Claude Security in public beta for enterprise customers Anthropic tools
11 мая Fake OpenAI Repo Hits #1 Trending on Hugging Face with 244K Downloads, Delivers Infostealer tools
25 июн Anthropic Accuses Alibaba of Largest Known Claude Distillation Attack: 28.8M Conversations Anthropic industry
26 авг Alabama AG subpoenas OpenAI over July AI-agent escape that hacked Hugging Face OpenAI industry
12 сен Researchers link OpenAI test agents to undisclosed May attack on RubyGems OpenAI tools
14 мая OpenClaw v2026.5.12-beta.4/5/6: Security Hardening and Multi-Platform Messaging Fixes tools
2 июн Claude Code v2.1.160: Security Prompts Before Writing Shell Startup Files and Build-Tool Configs Anthropic tools
4 июн Claude Code v2.1.162: Security Fix for OAuth Credential Leak, Parallel Tool Call Isolation Anthropic tools
6 июн Claude Code v2.1.166: Fallback Model Config, Expanded Deny-Rule Globs, Cross-Session Security Anthropic tools
24 июн Claude Code v2.1.187: Sandbox Credential Isolation and Remote MCP Hang Fix Anthropic tools
26 июн Claude Code v2.1.193: Shell Classifier Expansion, OTel Response Logging, Live Path Autocomplete Anthropic tools
26 июн OpenAI Codex CLI v0.142.2: Default MCP Tool Search, macOS Proxy Support, PowerShell Safety OpenAI tools
30 июн Claude Code v2.1.196: Org Default Models and MCP Security Fix for Untrusted Repos Anthropic tools
2 июл OpenAI Codex v0.142.5: Security Fix for Trace Log WebSocket Exposure OpenAI tools
4 июл GitHub Copilot CLI Drops PAT Requirement in Actions; Agent Session Streaming in Preview GitHub tools
9 июл Claude Code v2.1.205: Transcript-Tampering Guard and Session Reliability Fixes Anthropic tools
11 июл Claude Code v2.1.207: Auto Mode GA on Cloud Providers, Security Fix Anthropic tools
12 июл GitHub Copilot Adds Prompt-Injection Detection in CodeQL 2.26.0 GitHub tools
16 июл Claude Code v2.1.210–211: Bidirectional-Override Security Patch and Subagent Stream Flag Anthropic tools
19 июл Claude Code v2.1.214–v2.1.215: permission-bypass fixes, EndConversation tool, /verify no longer auto-runs Anthropic tools
3 авг Safeguards Based on Copyable Context Cannot Provide Reliable Safety for LLMs research
4 авг Claude Code v2.1.221 adds Focus view and sandbox credential masking Anthropic tools
6 авг Claude Code v2.1.223 fixes permission-bypass gap, merges /code-review Anthropic tools
13 авг Stealing Reasoning Traces from Proprietary LLM APIs research
14 авг Claude Code v2.1.232: subagent forking on by default, cross-session mentions, GitLab plugin support Anthropic tools
17 авг OpenClaw 2026.8.1-beta.2 pre-release ships GPT-5.6 Ultra switching and secret-egress binding tools
19 авг Claude Code v2.1.235 adds optional spellcheck, fixes window-restore focus and embedded grep Anthropic tools
25 авг Zed v1.16.2 — filesystem sandbox escape fix, GitHub Copilot Chat on GitHub Enterprise Cloud Zed tools
1 сен OpenAI Codex CLI 0.152.0 ships vim search, MCP output limits, credential-protection fix OpenAI tools
9 сен Meta details Muse agent safety architecture and opens its bug bounty to everyone Meta AI tools
11 сен Claude Code v2.1.268 fixes HTTP 400 regression and symlink permission bypass Anthropic tools
11 сен OpenClaw ships 2026.6.35, the final June 2026 LTS release tools
17 сен Cline Desktop v0.0.29 patches undici CVE-2026-1525 and enables web search by default Cline tools
17 сен llama.cpp patches remotely exploitable use-after-free in llama-server RPC endpoint ggml tools
18 сен Claude Code v2.1.275/276: claude.ai skill sync, ctrl+enter send-now, npm plugin hardening Anthropic tools
18 сен Codex CLI 0.155.0 goes stable with /voice, Touch ID MCP approvals, sandbox hardening OpenAI tools
18 сен Cline ships SSH remotes and concurrent sub-agents (Desktop 0.0.31/0.0.32, ext 4.1.19) Cline tools
26 июн OpenAI Ships codex-zsh v0.1.0: Versioned Patched zsh Binary for Codex Sandbox OpenAI tools
4 июл OpenAI Codex CLI 0.142.5: Trace-Log Privacy Fix OpenAI tools
5 авг Claude Code v2.1.222 fixes worktree destructive-git-command and hook-bypass safety gaps Anthropic tools
6 авг OpenAI Codex CLI 0.146.1 hardens auto-review for cyber-capable models OpenAI tools
18 сен Claude-Red: a curated offensive-security skills library gains 6k stars SnailSploit tools