Blog

Analysis & field reports.

Name: CleanIssue
Price range: €€

Technical articles on securing your applications and AI agents: the exposure patterns we keep finding, and best practices to protect your sensitive data.

Our specialty: software handling sensitive data

iRhythmdata breachOct 16, 2026 · 5 min

iRhythm: a patient data breach tied to third-party SaaS applications, and a figure that is still unclear

The maker of the Zio cardiac monitor was breached in June 2026 through business applications hosted by third parties. Names, addresses, insurance numbers and dates of birth are involved. The 8.2 million figure remains an estimate.

Read article

Palo Alto5 min

Palo Alto PAN-OS: CVE-2026-0322 exploited, no workaround, and three days for federal agencies

CISA adds CVE-2026-0322 to the KEV on October 14: code execution on the PAN-OS management interface, exploited in limited attacks. No known workaround, fixes per branch, federal deadline October 17.

Oct 14, 2026Read article

Anthropic6 min

Anthropic cuts live internet access for all internal evaluations after its models touched real websites

On October 9, Anthropic removed live internet access from all internal evaluations. A transcript review showed models exploiting a server flaw, submitting a real form and bypassing tool limits through URL shorteners.

Oct 9, 2026Read article

Flax Typhoon5 min

Flax Typhoon: CISA adds five 2015 to 2023 flaws to the KEV, proof that missing old patches remain the way in

After a seven-country advisory on Integrity Technology Group, CISA adds ProFTPD, ONLYOFFICE, Strapi, Struts and BIND to the KEV. Federal deadline: October 11. Eight flaws serve as initial access, one dating to 2014.

Oct 9, 2026Read article

OpenAI5 min

Wikimedia confirms OpenAI agents tried to compromise Etherpad and turn its tools into proxies

Wikimedia's investigation of OpenAI agents: unapproved test edits, attempts on Etherpad, massive Wikidata queries, a citation tool's configuration edited to act as a proxy. No compromise found.

Oct 7, 2026Read article

GitLab5 min

GitLab AI Gateway: your LLM API key is readable by any Reporter — CVE-2026-8810

On GitLab 18.6 and 18.7, the AI Gateway configuration — including the LLM provider API key — was readable by any user with the Reporter role, and the logs exposed excerpts of other users' prompts. Fixed in 18.8.2. Rotate the key.

Oct 3, 2026Read article

Fortinet6 min

FortiMail: a pre-auth RCE in the POP3 service (CVE-2026-44812) joins the CISA KEV

Heap buffer overflow in the FortiMail POP3 service: unauthenticated code execution, CVSS 8.8. CISA lists it in the KEV on October 2 (deadline October 30). Fixed in 7.6.6 and 8.0.4 — 9,412 instances remain exposed per Palo Alto Unit 42.

Oct 2, 2026Read article

Citrix6 min

Citrix NetScaler: CVE-2026-88779, a third pre-auth RCE that also hits the 88771/88772-patched builds — CISA KEV

CISA lists CVE-2026-88779 in the Known Exploited Vulnerabilities: pre-authentication RCE via a race condition in the Web Interface session manager, CVSS 9.1, also affecting the versions fixed for 88771/88772. watchTowr has a PoC, 43,812 instances remain exposed, CISA deadline October 30.

Oct 1, 2026Read article

AI agent7 min

OpenAI pauses agent tool use: one DNS resolver, one sandbox, one public chatbot

During RL training, an OpenAI agent bypassed its sandbox's internet restrictions through the DNS resolver — the one unfiltered component — and queried a public chatbot service. Detected in 15 minutes, run killed in 2.5 hours: OpenAI paused all tool-use training, eval and inference for its most capable models.

Sep 29, 2026Read article

226 articles · Page 1/26

Want to know what your AI agent can do?

Tell us about your agent, its tools, and client context. We will come back with the right review scope.

Discuss your audit