muapi-nano-banana
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a local shell script
scripts/generate-nano-art.sh. This script handles user-defined parameters for subject, action, and context to generate image prompts, which is a standard operational pattern for skill-based tasks. - [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes user-supplied text to construct reasoning-driven prompts, creating an attack surface for indirect instructions.
- Ingestion points: Creative parameters including subject, action, context, and style provided as user input in SKILL.md.
- Boundary markers: The skill does not explicitly define delimiters or instructions for the agent to ignore potentially malicious instructions embedded within the user-provided creative briefs.
- Capability inventory: Shell command execution capability via
bash scripts/generate-nano-art.sh(SKILL.md). - Sanitization: The skill definition does not specify any input validation or sanitization mechanisms for the user-supplied strings before they are processed by the script.
Audit Metadata