trivy

Warn

Audited by Snyk on Feb 19, 2026

Risk Level: MEDIUM
Full Analysis

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

  • Third-party content exposure detected (high risk: 0.90). The skill's workflow explicitly shows commands like "trivy repo https://github.com/owner/repo" and other "repo" and remote-scan examples that fetch and analyze arbitrary public git repositories (user-generated content) which the agent would read and whose contents could influence scanning decisions and follow-up actions.

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

  • Attempt to modify system services in skill instructions detected (high risk: 0.70). The prompt includes explicit installation steps that use sudo and commands that write to system locations (e.g., /usr/share/keyrings, /usr/local/bin, /etc/apt/sources.list.d), which instruct modifying the host system and require elevated privileges, so it can push an agent to change the machine state.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 19, 2026, 07:52 AM