Service Creator

Pass

Audited by Gen Agent Trust Hub on Mar 5, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill features an indirect prompt injection surface as it is designed to ingest data from user-defined external APIs and pass it to an LLM for evaluation via the invoke tool.
  • Ingestion points: Service scripts (e.g., reference/templates-stock-monitor.md) fetch data from external URLs.
  • Boundary markers: No explicit instructions are provided to use delimiters or 'ignore' instructions when interpolating external data into prompts.
  • Capability inventory: The system executes these scripts persistently via the service_start tool.
  • Sanitization: Guidelines emphasize response structure validation but do not specify content-based sanitization to mitigate instruction injection.
  • [COMMAND_EXECUTION]: The skill uses dynamic script generation and system command execution as part of its core functionality.
  • Evidence: The workflow utilizes service_create and service_start to write and execute code on the host system.
  • Evidence: The instructions suggest using package managers like Homebrew to install runtimes (e.g., 'brew install node').
  • Context: These actions are necessary for the skill's primary purpose and are mitigated by constraints against third-party dependencies and elevated privileges.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 5, 2026, 07:26 AM