ui-slot-ux-designer

Pass

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a local Python script scripts/validate_slot_ux_spec.py to validate JSON-based UX specifications. This is the primary function of the skill and is implemented using standard local execution.\n- [PROMPT_INJECTION]: There is a potential surface for indirect prompt injection through the processing of untrusted input data.\n
  • Ingestion points: The validate_slot_ux_spec.py script reads an external JSON specification file provided by the user via the --input flag.\n
  • Boundary markers: The instructions do not define clear delimiters or include warnings for the agent to disregard natural language instructions that might be present in the data fields of the JSON file.\n
  • Capability inventory: The skill has the capability to read files and generate structured patch plans and UX maps based on input content.\n
  • Sanitization: While the script validates the JSON schema and data types, it does not sanitize string fields for potential natural language instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 10, 2026, 03:29 AM