muapi-blog-header
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user data through fields like topic and publication_style, which are directly interpolated into the final prompt for the image generation model.\n
- Ingestion points: Input variables in SKILL.md (topic, publication_style, dominant_color).\n
- Boundary markers: Absent.\n
- Capability inventory: Command execution via CLI and network access via curl.\n
- Sanitization: Absent.\n- [COMMAND_EXECUTION]: The skill instructs the agent to perform command-line operations using the muapi CLI tool and curl to send data to the muapi.ai API. This represents standard operational behavior for this skill's functionality.
Audit Metadata