eto-apere
Pass
Audited by Gen Agent Trust Hub on Aug 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions require the agent to ingest and analyze existing project files, including token definitions, theme files, and component conventions. This creates an attack surface where malicious instructions embedded in these files could influence agent behavior.\n
- Ingestion points: Project-native token files, theme files, and design system documentation (SKILL.md).\n
- Boundary markers: Absent; no specific instructions are provided to the agent to disregard or delimit embedded instructions in the source files.\n
- Capability inventory: The skill utilizes a bundled Node.js script (scripts/generate-tokens.cjs) which can be executed via the command line and has the capability to write to the filesystem.\n
- Sanitization: Absent; input files are parsed as JSON but not sanitized for prompt-based instructions.\n- [COMMAND_EXECUTION]: The skill provides and instructs the use of a bundled Node.js script (scripts/generate-tokens.cjs) to process design tokens and generate CSS or Tailwind configurations. This involves local shell execution to perform its primary function.
Audit Metadata