theme-factory

Pass

Audited by Gen Agent Trust Hub on Mar 6, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION]: The skill instructions for generating custom themes and styling external artifacts present an indirect prompt injection surface.
  • Ingestion points: User-provided theme descriptions and untrusted artifacts (slides, docs, HTML) as specified in SKILL.md.
  • Boundary markers: No delimiters or instructions to ignore embedded commands within processed data are present.
  • Capability inventory: The skill relies on agent capabilities to read from and write to file systems to apply theme changes.
  • Sanitization: No input validation or content sanitization mechanisms are defined, although the skill includes a human review step for custom themes.
  • [NO_CODE]: The skill contains no executable scripts, binaries, or complex automation, relying solely on natural language instructions and markdown configuration files.
  • [SAFE]: No hardcoded credentials, malicious network operations, or obfuscated content were detected in the theme definition files or instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 6, 2026, 06:39 AM