muapi-multi-angle-shots
Pass
Audited by Gen Agent Trust Hub on May 19, 2026
Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it takes untrusted user input (e.g., product_name, background_style) and embeds it directly into prompts for image generation models.
- Ingestion points: Input variables defined in the SKILL.md frontmatter (product_name, background_style, lighting, category).
- Boundary markers: None present to distinguish user input from instructions within the generated prompts.
- Capability inventory: Use of muapi image generate and curl network operations to external services.
- Sanitization: No validation or escaping of the user-provided text is performed before interpolation.
- [DATA_EXFILTRATION]: The instructions direct the agent to transmit a sensitive API key to an external endpoint (api.muapi.ai) using curl if the primary CLI tool is unavailable.
- [COMMAND_EXECUTION]: The skill relies on executing shell commands, specifically the muapi CLI tool for image generation and task management, and curl for fallback API interactions.
Audit Metadata