muapi-ai-fight-scene

Pass

Audited by Gen Agent Trust Hub on May 19, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to prompt injection because it interpolates untrusted user inputs (character_description, environment_description, action_script, style_direction) directly into model prompts across all phases without boundary markers or instructions to disregard embedded commands.
  • Ingestion points: Inputs are gathered via the Inputs table and used in the Steps section of SKILL.md.
  • Boundary markers: Absent. User text is placed directly within the prompt strings.
  • Capability inventory: The skill can perform network operations (via curl) and invoke external CLI tools (muapi) to generate images and videos.
  • Sanitization: None. There is no evidence of escaping or validation of the user-provided strings before they are sent to the models or CLI.
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using the muapi CLI and curl. While these are part of the core functionality, the use of $MUAPI_API_KEY and the execution of commands containing user-provided text require the agent to handle environment variables and string interpolation securely to prevent unintended shell behavior.
Audit Metadata
Risk Level
SAFE
Analyzed
May 19, 2026, 02:04 PM
Security Audit — agent-trust-hub — muapi-ai-fight-scene