movie-poster-graphics

Warn

Audited by Gen Agent Trust Hub on Oct 2, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill recommends using npx minibanana@0.1.0 as a wrapper for OpenRouter access. This command downloads and executes a package from the public npm registry at runtime. The package is not published by the skill author (skillatlas) or a recognized trusted organization.
  • [COMMAND_EXECUTION]: The skill provides specific shell command patterns for executing external packages to perform network operations and image generation tasks.
  • [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it incorporates user-derived concepts into the prompts sent to AI image models.
  • Ingestion points: User requests are used to generate [CONCEPT_DESCRIPTION], [TITLE], and [TAGLINE] variables in SKILL.md.
  • Boundary markers: The skill uses quotes around title and tagline text in its final prompt formula, which provides only minor protection against adversarial input.
  • Capability inventory: The skill possesses shell execution capabilities and network access to third-party image generation APIs.
  • Sanitization: There are no instructions for sanitizing or filtering user-provided concept descriptions before they are interpolated into the final prompt.
  • [DYNAMIC_EXECUTION]: The skill involves generating prompt files (e.g., @poster-prompt.md) at runtime to be used as input for the minibanana execution tool.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Oct 2, 2026, 12:55 PM