hypervisor-introspection

Warn

Audited by Snyk on Nov 14, 2026

Risk Level: MEDIUM
Full Analysis

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

  • Third-party content exposure detected (low risk: 0.10). The required workflow is mostly local CLI enumeration and lab reproduction, but the payloads guidance tells the agent to search GitHub for a public VENOM PoC, which means ingesting arbitrary outsider-authored repositories it must actively select (low, not medium, since no outsider text is pushed into a monitored queue).

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

  • Potentially malicious external URL detected (medium risk: 0.30). The decisive boundary is the instructed setup path that clones and builds code from an individual-owned repository (tandasat/HyperPlatform, which is then loaded as a kernel driver), so this is a common/reputable unofficial dependency and the highest tier is medium, while the QEMU, PANDA, and Bitdefender KVMi clones are strong/official.

MEDIUM W013: Attempt to modify system services in skill instructions.

  • Attempt to modify system services in skill instructions detected (medium risk: 0.40). The skill is a security-knowledge reference, but it tells the agent to install kernel-mode drivers and hypervisors on a host, write system config under /etc, and stop system services, which are machine-state changes that can require elevated privileges.

Issues (3)

W011
MEDIUM

Third-party content exposure detected (indirect prompt injection risk).

W012
MEDIUM

Unverifiable external dependency detected (runtime URL that controls agent).

W013
MEDIUM

Attempt to modify system services in skill instructions.

Audit Metadata
Risk Level
MEDIUM
Analyzed
Nov 14, 2026, 11:21 AM
Issues
3