ZeroDay Findings
Sign inSign up

Every 4 hours

Cyber

Original briefings from the security press, rewritten for operators. Each story keeps the date of the report it is based on, links the identifiers it discusses, and points back to the original publication.

Threat intelligence

Trackers and reporting desks followed on the same four-hour update. New public reports are rewritten into the briefings below.

PrioritySourceWebsiteInformation
Ransomware.liveWebsiteRansomware.liveInformationRansomware groups, victims, incidents
The DFIR ReportWebsiteThe DFIR ReportInformationDetailed intrusion investigations
SANS Internet Storm CenterWebsiteSANS ISCInformationAttack observations and incident analysis
abuse.ch ThreatFoxWebsiteThreatFoxInformationIndicators of compromise
abuse.ch URLhausWebsiteURLhausInformationMalware distribution infrastructure
abuse.ch MalwareBazaarWebsiteMalwareBazaarInformationMalware hashes and metadata
abuse.ch Feodo TrackerWebsiteFeodo TrackerInformationBotnet command-and-control infrastructure
AlienVault OTXWebsiteAlienVault OTXInformationCommunity threat-intelligence indicators
PhishTankWebsitePhishTankInformationPhishing intelligence
Shadowserver FoundationWebsiteShadowserverInformationInternet exposure and threat reporting

  1. Oct 16, 2026 · CSO Online

    Supply Chain Vulnerability Discovered in OpenAI Skills Repository

    A newly identified supply chain vulnerability in OpenAI's skills repository raises concerns about AI security and trust in vetted components.

    What changed

    Researchers have uncovered a significant supply chain vulnerability within OpenAI's official skills repository. This flaw involved a skill that directed applications to send sensitive tokens to an unregistered domain, potentially compromising production environments.

    • The vulnerability was introduced with a commit made in January.
    • The affected skill was designed to facilitate application deployment on Render.
    • The domain render-internal.com was registered by researchers to demonstrate the flaw.

    Who is affected

    The impact of this vulnerability extends to several stakeholders within the AI and cloud application ecosystem. Organizations utilizing OpenAI's skills for application deployment may be at risk.

    • OpenAI, as the maintainer of the skills repository.
    • Render, the platform for which the flawed skill was designed.
    • Developers and companies using OpenAI's skills to integrate AI into their applications.

    What to do

    Organizations should review their use of OpenAI skills and ensure that they are not inadvertently exposing sensitive information. It is crucial to stay informed about updates and patches related to this vulnerability.

    • Audit applications for any references to render-internal.com.
    • Ensure that the OpenAI skills being used are the latest versions post-fix.
    • Monitor for any unusual API calls or token transmissions from production applications.

    Original report