product-data-cleanup

Pass

Audited by Gen Agent Trust Hub on May 11, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes data from external, untrusted sources without implementing specific security boundaries or sanitization.
  • Ingestion points: Data is loaded from external Google Sheets using mcp__google-sheets__get_sheet_data and local files using the Read tool as described in the SKILL.md input section.
  • Boundary markers: The instructions do not specify the use of delimiters or provide the agent with explicit instructions to ignore prompt-like content found within the data rows during the cleaning process.
  • Capability inventory: The skill environment includes powerful tools such as Bash, Write, and mcp__google-sheets__update_cells, which create a significant attack surface if an injection is successfully processed.
  • Sanitization: While the skill provides extensive rules for formatting and normalizing data (casing, dimensions, currency), it does not mention sanitizing input to prevent the execution of embedded instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
May 11, 2026, 11:58 AM
Security Audit — agent-trust-hub — product-data-cleanup