cve-triage

Pass

Audited by Gen Agent Trust Hub on Jul 1, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill is designed around executing various shell commands via the infynon pkg CLI, including scanning for vulnerabilities, performing automated fixes, and auditing dependency trees.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8). It instructs the agent to ingest and parse JSON output from infynon pkg scan and directly execute remediation steps based on the fix_cmd field provided in that data.
  • Ingestion points: The agent reads and processes output from the infynon pkg scan --json command.
  • Boundary markers: There are no explicit markers or instructions to treat the ingested JSON data as untrusted or to ignore instructions embedded within it.
  • Capability inventory: The agent has shell access via the infynon pkg command, which allows for package installation, environment modification, and system auditing.
  • Sanitization: The instructions do not specify any validation or sanitization of the fix_cmd or package fields before interpolation into execution commands.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 1, 2026, 07:15 AM
Security Audit — agent-trust-hub — cve-triage