Claude, Codex, and Hermes installed unowned code inside corporate networks
Source: Ars Technica (opens in a new tab) · Dan Goodin
Intel Summary
Security reporting indicates that coding models including Claude, Codex, and Hermes have generated package installation commands referencing unowned software code inside corporate documentation. An analysis identified 227 install commands pointing to unregistered package names. This hallucination phenomenon creates software supply chain risks where threat actors can register the vacant package names on public registries, enabling potential remote code execution or data theft inside corporate networks.
Why It Matters
AI package hallucination represents a critical software supply chain vulnerability for enterprise software development. When developers or automated AI agents execute AI-suggested install commands without strict package validation, malicious actors can exploit the gap via namespace registration. Organizations must implement dependency scanning, package verification, and private registry guardrails to mitigate risks from automated AI code generation.
Part of an ongoing development
Independent reportingClaude, Codex, and Hermes generate unowned package installation commands
Security reporting indicates that coding models including Claude, Codex, and Hermes have generated package installation commands referencing unowned software code inside corporate documentation. An analysis identified 227 install commands pointing to unregistered package names. Claims are as reported; this summary makes no determination about accuracy or significance.
- Confidence
- Moderate confidence
- Corroboration
- Limited corroboration
Organizations & Entities
Related Intelligence
- ReportAlso involving Claude
Salesforce Customers Brace For New Fees to Use Claude in Its Apps
Salesforce plans to require enterprise customers to register external AI agents, such as Anthropic's Claude, beginning as early as next month, according to The Information. The registration process serves as an initial step toward charging fees whenever third-party AI agents update or analyze data stored inside Salesforce apps, allowing the vendor to monetize users who deploy external AI rather than its proprietary tools.
The Information - DevelopmentDevelopingAlso involving Claude
Mistral unveils new AI model
Mistral has unveiled a new AI model that the company claims rivals leading open AI systems from China, according to CNBC. The release comes amid intensifying competition between Western and Chinese developers in the open-weight and open-source model space. Claims are as reported; this summary makes no determination about accuracy or significance.
3 independent sources - ReportAlso involving Claude
Google Cloud now offers OpenAI models alongside Gemini and Claude, becoming the only major provider with all three
Google Cloud has begun offering OpenAI models alongside Gemini and Claude, making it the first major hyperscaler to host all three leading model families on one platform. At launch, GPT-6 Astra, GPT-6.1 Sol, and GPT-6 Luna are accessible in private preview for enterprise customers.
The Decoder - ReportAlso involving Claude
Hackers Used Claude and Grok to Fully Automate Attacks on Banks
According to an investigation by cybersecurity startup Straiker reported by The Information, a hacking group named CyberCartel deployed a network of AI agents powered by models from Anthropic, SpaceX, and Moonshot AI to automate cyberattacks against more than 400 Latin American financial targets. The research represents one of the first documented cases of threat actors using multi-model agentic AI systems to successfully automate intrusions into financial institutions.
The Information