token-cost-ranking

Pass

Audited by Gen Agent Trust Hub on Jul 28, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a bash command dce business-cockpit ... --hostname <host> using a user-supplied <host> parameter. Direct interpolation of user-controlled strings into shell environments presents a command injection risk if the underlying agent platform does not perform strict sanitization.
  • [PROMPT_INJECTION]: The skill processes data from external tool output (Category 8: Indirect Prompt Injection). There is a risk that malicious data returned by the dce command (e.g., within the tenantName field) could attempt to manipulate the agent's subsequent responses.
  • Ingestion points: JSON items returned by the dce business-cockpit command.
  • Boundary markers: None specified for the tool output.
  • Capability inventory: Shell command execution (dce CLI).
  • Sanitization: The skill includes logic to validate that the price field is numeric before processing, which provides some protection for that specific field.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 28, 2026, 08:52 AM
Security Audit — agent-trust-hub — token-cost-ranking