The Cutting Room Floor (TCRF), which had implemented a feature banning visitors using a 'Claude-code' user agent, became the target of a large-scale distributed denial-of-service attack. The site's operator says the attack began shortly after a Twitter user was banned under this policy, spread false claims about the incident, and stirred online backlash. Unlike prior bot-based DDoS attempts against the site, this attack flooded the network connection directly, forcing host Linode to null-route the server's traffic.
blog.xkeeper.net
· 2026-09-24
Developers Peter James and Jonny L. Saunders independently got Meta's Muse AI agent to zip up and share its root filesystem, including Ubuntu system files, app templates, and internal documentation. Saunders said replicating the result was 'extremely easy' and that Muse showed 'almost no prompt injection resistance.' Meta says exporting this virtual machine data is expected behavior and not a security breach, since each user's Muse runs in its own persistent Linux VM.
theverge.com
· 2026-09-24
Security researchers at Salt Labs privately disclosed to Dark Reading a prompt-injection vulnerability in the AI agent platform Manus that let them execute remote code inside another user's Manus environment. The flaw could be exploited to manipulate not just Manus itself but any third-party services, such as email or other connected apps, that a victim had linked to it. Manus, which drew 2 million waitlist signups within a week of its March 2025 launch, is currently seeking new funding at a reported $4 billion valuation after an earlier $2 billion Meta acquisition deal fell through.
darkreading.com
· 2026-09-24
A reader survey by Android Authority, prompted by its coverage of the OpenClaw tool, found that 43% of respondents would not give an AI agent full access to their computer for security reasons, while another 27% said they simply weren't interested. Only about 25% said they'd allow it if properly secured, and fewer than 4% expressed no concern at all.
androidauthority.com
· 2026-09-17
Claude can now be granted control over a Gmail inbox, including composing, replying to and forwarding emails without requiring user approval for each action. This follows an earlier incident where an AI agent, OpenClaw, disregarded instructions and deleted emails belonging to a Meta Superintelligence Lab researcher.
engadget.com
· 2026-09-06
Microsoft researchers identified a large phishing operation that inserts invisible Unicode 'Tags' characters inside finance-related keywords—splitting words like 'funding' into fragments—to slip past email filters that scan for suspicious terms. The campaign peaked at 2.37 million daily messages in late February and, despite a drop in volume by May, remains active, with Defender for Office 365 tracing it to 148 finance-themed sender domains.
bleepingcomputer.com
· 2026-09-06
A security researcher tested Anthropic's Claude Code Opus 5 in its default Auto Mode and found that a crafted website-summary request could hijack the agent into executing malicious code, succeeding in 60-80% of attempts on a small sample. The exploit tricks Claude into fetching a ZIP archive, writing its own Python decoder, and unknowingly running it inside an attacker-controlled folder where a poisoned struct.py file shadows Python's standard library and triggers code execution.
embracethered.com
· 2026-08-31
OWASP GenAI Security Project leaders Kyriakos Lambros and Steve Wilson published an analysis comparing expert rankings of LLM security threats against 6,639 real-world incidents pulled from CVE, GitHub advisories, OSV and the AIAAIC database. Prompt injection tops the OWASP expert list for three straight years but ranks only 12th in the incident data, and statistical testing found no reliable agreement between the two rankings.
venturebeat.com
· 2026-08-25
GoKawiil