First reported embracethered.com
Lead dispatch
First reported island.io
AgentBaiting: How Fake AI Skills Deliver Malware at Scale
Island security researchers uncovered the FakeGit campaign, which used roughly 7,600 malicious GitHub repositories—over 800 posing as AI Skills or MCP servers—to deliver SmartLoader malware that installs the StealC information stealer. The campaign introduces a technique called AgentBaiting, in which AI agents such as Claude Code, Gemini, and ChatGPT autonomously discover the attacker repositories, treat the malicious README as legitimate documentation, and pass installation instructions to users; the operation recorded over 14 million downloads.supply-chain · tool-abuse · malware-delivery · agentbaiting
mcp · ai-agents · ai-skills · llm · copilot
Severity
0.75
The wire · latest
First reported pillar.security
The Week of Sandbox Escapes
Pillar Security researchers demonstrated seven sandbox-escape techniques against four AI coding agents — Cursor, OpenAI's Codex, Google's Gemini CLI and Antigravity — where a sandboxed agent writes workspace files that trusted tools running outside the sandbox (Python extensions, Git integrations, hook engines, Docker daemons) later execute. Prompt injection planted in a README, issue, dependency, or diff triggers local command execution on the developer's machine; one Cursor flaw is tracked as CVE-2026-48124 and fixed in version 3.0.0, with most issues patched and vendor-acknowledged. Details →First reported · updated · 5 reports wiz.io
GhostApproval: AI Coding Assistant Trust Boundary Flaw
Wiz researchers disclosed GhostApproval (CVE-2026-12958 and related), a symlink-following vulnerability pattern (CWE-61 plus CWE-451 UI misrepresentation) affecting six top AI coding assistants: Amazon Q Developer, Anthropic Claude Code, Augment, Cursor, Google Antigravity, and Windsurf. A malicious repository disguises a symlink as an innocuous config file so that when the agent 'sets up the workspace' it writes an attacker SSH key to ~/.ssh/authorized_keys — sometimes before any confirmation dialog — while the human-in-the-loop approval prompt conceals the true target, enabling remote code execution. AWS, Cursor, and Google fixed the issue; two vendors went silent and one called it outside its threat model. Details →First reported · updated · 5 reports varonis.com
SearchLeak: How We Turned M365 Copilot Into a One-Click Data Exfiltration Weapon
SearchLeak (CVE-2026-42824) is a critical three-stage vulnerability chain in Microsoft 365 Copilot Enterprise discovered by Varonis Threat Labs that lets an attacker steal MFA codes, emails, meeting details, and organizational files with a single click on a trusted microsoft.com link. It chains a Parameter-to-Prompt (P2P) injection via the search q parameter with an HTML rendering race condition and a CSP bypass through Bing's allowlisted image-search SSRF endpoint to silently exfiltrate a victim's mailbox, calendar, SharePoint, and OneDrive data. Microsoft remediated the flaw and rated it critical. Details →First reported wiz.io
Wiz Uncovers Critical Access Bypass Flaw in AI-Powered Vibe Coding Platform Base44
Wiz disclosed a now-patched critical authentication bypass in Base44, an AI-powered 'vibe coding' platform owned by Wix, where supplying only a non-secret 'app_id' value to undocumented registration and email-verification endpoints let an attacker create a verified account for private applications. The flaw bypassed all authentication controls, including SSO, granting full access to private apps and their data; Wix fixed it within 24 hours of the July 9, 2025 disclosure, with no evidence of exploitation in the wild. Details →First reported · updated · 2 reports cyera.com
Ollama Out-of-Bounds Read Vulnerability Allows Remote Process Memory Leak
Cyera researchers disclosed CVE-2026-7482 ("Bleeding Llama", CVSS 9.1), a heap out-of-bounds read in Ollama's GGUF model loader that lets a remote, unauthenticated attacker leak the entire process memory of an Ollama server. The flaw is fixed in Ollama 0.17.1 and is estimated to affect over 300,000 exposed servers globally. Details →First reported strobes.co
CVE-2026-23744 - CVE Details, Severity, and Analysis | Strobes VI
CVE-2026-23744 is a critical (CVSS 9.8) remote code execution vulnerability in MCPJam Inspector versions 1.4.2 and earlier, a local-first development platform for MCP servers. Because the tool binds to 0.0.0.0 and its /api/mcp/connect endpoint extracts command and args without security checks, an attacker can send a crafted, unauthenticated HTTP request to trigger arbitrary command execution with no user interaction; public PoCs and vendor patches are available. Details →First reported nist.gov
NVD - CVE-2026-59819
CVE-2026-59819 is a vulnerability in LiteLLM, an AI gateway/proxy for calling LLM APIs, where prior to version 1.83.10-stable the /health/test_connection endpoint resolved request-supplied environment and OIDC file references in litellm_params, allowing a privileged caller to read arbitrary local files via an oidc/file/ reference (CWE-73). The issue is fixed in version 1.83.10-stable. Details →First reported · updated · 3 reports theregister.com
Red teamers turned Claude Desktop into a double agent to do their evil bidding
Oasis Security disclosed "PromptFiction," a vulnerability in Anthropic's Claude Desktop where a single click on a trusted-looking claude:// URL silently submitted attacker-controlled prompts to the assistant with no user confirmation. Chained with the earlier "Claudy Day" trio of flaws, it could enable silent exfiltration of prior conversations and — when Anthropic's official Filesystem MCP server is installed — file read/write, persistence, and remote code execution. Anthropic has fixed the flaw. Details →First reported embracethered.com
From Indirect Prompt Injection to DNS Exfiltration in macOS Terminal
Johann Rehberger (Embrace The Red) demonstrates a proof-of-concept where an indirect prompt injection hidden in a spreadsheet cell hijacks an LLM-integrated CLI tool to emit ANSI escape codes, which the macOS Terminal app interprets to issue DNS requests that exfiltrate data from other rows. Building on prior Terminal DiLLMa research and David Leadbeater's original finding, the post reports that Apple has fixed the macOS Terminal behavior enabling the DNS-based exfiltration. Details →First reported · updated · 2 reports paloaltonetworks.com
Pickle in the Middle – Hijacking Vertex AI Model Uploads for Cross-Tenant RCE
Palo Alto Unit 42 (Ori Hadad) disclosed CVE-2026-0257, a flaw in Google's Vertex AI Python SDK (python-aiplatform) where model uploads relied on an attacker-predictable Google Cloud Storage bucket, enabling 'bucket squatting.' An attacker could pre-register the expected bucket and substitute a malicious pickle/joblib model artifact, achieving cross-tenant remote code execution when the poisoned model was deserialized. Google fixed the issue in SDK releases v1.148.0/v1.148.1. Details →First reported bugcrowd.com
One-Click Data Exfiltration via rovoChatPrompt URL Parameter (Confluence / Rovo) - CrowdStream
A disclosed and now-patched vulnerability in Atlassian Rovo, the default AI assistant across Confluence and other Atlassian products, allowed the `rovoChatPrompt` URL parameter to preload an arbitrary prompt into a victim's Rovo chat. When an authenticated user clicked a crafted link, Rovo executed the embedded prompt as a genuine query, using indirect task-framed language ('help me identify this image') to bypass guardrails and exfiltrate Confluence pages, secrets, and connected-app data (Jira, SharePoint, Outlook) to an attacker host via an image-fetch URL. Atlassian remediated the issue server-side and the reporter validated the fix. Details →First reported · updated · 3 reports catonetworks.com
DuneSlide: Two Critical RCE vulnerabilities
Cato AI Labs disclosed two critical vulnerabilities in Cursor IDE, dubbed DuneSlide (CVE-2026-50548 and CVE-2026-50549, both CVSS 9.8), that allow zero-click prompt injection to escape the editor's command sandbox and achieve remote code execution. CVE-2026-50548 abuses the run_terminal_cmd tool's working_directory parameter to add attacker-chosen folders to the allowed-write list, letting an attacker overwrite the cursorsandbox binary or shell startup files and turn sandboxed commands into full system compromise. Cursor patched both in version 3.0, released April 2; earlier 2.x builds remain exposed. Details →First reported · updated · 2 reports cereblab.com
Grok Build Uploads Entire Git Repositories to xAI Storage, Not Just Files It Reads
AI safety researcher Cereblab found that xAI's Grok Build CLI uploaded users' entire Git repositories — including files it was explicitly told not to open and secrets deleted from Git history — as Git bundles to a Google Cloud Storage bucket, even on an idle 'reply OK' prompt. The behavior was reproduced across multiple codebases and captured on the wire via mitmproxy; xAI later disabled it server-side with a disable_codebase_upload flag and Musk promised deletion of uploaded data. Details →First reported jscrambler.com
Unauthorized Publication of a Malicious npm Package affecting CI
Jscrambler disclosed that a threat actor used compromised npm publishing credentials to publish a malicious version of its 'jscrambler' npm package (versions 8.14, 8.16, 8.17, 8.20), which ran an infostealer during the preinstall hook and was downloaded around 1,479 times in a two-hour window. Per Socket's analysis, the malware harvested developer credentials, cloud secrets, crypto wallets, browser data, and notably AI coding tool and MCP configurations (Claude, Cursor, Windsurf, VS Code, Zed), using ChaCha20-Poly1305 obfuscation to resist analysis. Jscrambler revoked credentials and released a safe version 8.22. Details →First reported arxiv.org
A Security Analysis of the OpenClaw AI Agent Framework
Researchers detailed three now-patched high-severity flaws (GHSA-hjr6-g723-hmfm, GHSA-9969-8g9h-rxwm, GHSA-575v-8hfq-m3mc; CVE-2026-46817, CVE-2026-55200) in the OpenClaw personal AI assistant framework that compose into a complete unauthenticated remote code execution path — from an LLM tool call triggered via a WhatsApp message to command execution on the host. The chain abuses OS command injection and an incomplete disallowed-input filter in the exec allowlist, whose closed-world lexical parsing is defeated by shell line continuation, busybox multiplexing, and GNU option abbreviation; a companion arXiv analysis taxonomizes 470 advisories and shows a malicious plugin skill bypassing the exec pipeline entirely. Details →First reported tabstack.ai
Hardening AI Web Agents: How We're Securing Tabstack Against Indirect Prompt Injection | Tabstack Blog
Tabstack (Mozilla) disclosed an indirect prompt injection vulnerability in its /v1/automate web-agent endpoint, found by Brave researchers, where hidden white-on-white instructions on a target page caused the autonomous agent to exfiltrate the user's conversation history to an external form. The flaw has been patched with a structural 'action firewall' that classifies form interactions via DOM metadata and provenance rather than trying to detect malicious text. Details →First reported microsoft.com
AutoJack: How a single page can RCE the host running your AI agent | Microsoft Security Blog
Microsoft researchers disclosed AutoJack, an exploit chain in AutoGen Studio where untrusted web content rendered by a browsing agent can reach a local MCP WebSocket and spawn arbitrary processes on the host, crossing the localhost trust boundary to achieve RCE. The issue was reported to MSRC and hardened upstream in commit b047730; the vulnerable MCP WebSocket surface was never shipped in a PyPI release. Details →How the wire is made
Poll & cluster
Internet is crawled for AI security news and near-duplicate coverage is embedded and grouped into durable items.
Curate
AI Agent filters for agentic-AI relevance, classifies and tags each item, scores severity for threats, and writes the summary.
Every item here is one machine-curated intelligence object, not a headline.
Read the wire for free. There is a small charge to ask the index questions.
The wire, open
The complete curated feed, no key required.
- GET /feed.xml — RSS 2.0, every item
- GET /api/items — read-only
The vector desk
Query the index by meaning, not just keyword.
- GET /api/items?tags=&minSeverity=&itemType=
- GET /api/search?q= — keyword
- GET /api/semantic?q= — vector