web-xxe
Audited by Socket on Sep 19, 2026
4 alerts found:
Securityx2Malwarex2SUSPICIOUS. This skill is an offensive exploitation playbook for XXE that instructs an AI agent to steal files, exfiltrate data, reach internal services, and weaponize Office documents for phishing. Its capabilities are coherent with its stated purpose, but that purpose is itself high-risk for AI-agent use because it operationalizes exploit and exfiltration workflows with real-world impact.
The visible fragment is an offensive XXE exploitation guide containing payloads and helper scripts for local-file disclosure, SSRF, cloud metadata theft, internal service probing, and out-of-band exfiltration. It does not by itself demonstrate hidden package malware or automatic execution, but it contains substantial dual-use attack tooling and explicit exfiltration logic. Treat the dependency as high risk unless it is intentionally a security-testing package and its execution paths are independently verified.
The fragment contains clear offensive XXE exploitation and data-exfiltration functionality. It includes an anonymous writable FTP collection server, dynamic DTD generation, local-file entity reads, attacker-controlled callback URLs, and WAF-bypass payload generation. This is highly risky if executed or deployed outside an authorized penetration-testing environment.
The supplied material is a high-risk offensive XXE exploitation and social-engineering guide. It contains concrete payloads and commands for weaponizing Office documents, reading local files through vulnerable XML parsers, exfiltrating data, uploading malicious documents, and monitoring callbacks. It is not obfuscated and is not itself a persistent malware implant, but it substantially facilitates unauthorized exploitation and data theft.