model-selection-advisor

Pass

Audited by Gen Agent Trust Hub on Jun 29, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill's instructions focus on model selection criteria and routing strategies using standard natural language. No patterns indicating safety filter bypass, system prompt extraction, or instruction override were found.
  • [DATA_EXPOSURE]: The skill does not access sensitive file paths, environment variables, or hardcoded credentials. It lacks any network capabilities for data exfiltration.
  • [REMOTE_CODE_EXECUTION]: There are no commands for downloading external scripts or installing third-party packages. The skill only references internal markdown documentation.
  • [INDIRECT_PROMPT_INJECTION]: While the skill ingests user-provided task descriptions, it does not possess any exploitable capabilities such as subprocess execution, file writing, or network operations that could be triggered by malicious inputs.
  • [DYNAMIC_EXECUTION]: The skill does not use dynamic code generation, runtime compilation, or unsafe deserialization of data.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 29, 2026, 10:26 PM
Security Audit — agent-trust-hub — model-selection-advisor