fantasy-movie-poster
Pass
Audited by Gen Agent Trust Hub on Jul 28, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection via the
scripts/build_prompt.pyscript.- Ingestion points: User-provided data such as
title,logline, andsynopsisare read from a JSON project file. - Boundary markers: There are no delimiters or boundary markers used when interpolating these strings into the final prompt template.
- Capability inventory: The agent is instructed to execute local Python scripts (
scripts/build_prompt.py,scripts/composite_layers.py) which handle file I/O and image manipulation. - Sanitization: There is no evidence of input validation, escaping, or sanitization before the external content is included in the prompt, allowing a user to potentially inject instructions that override the poster design constraints.
- Ingestion points: User-provided data such as
- [COMMAND_EXECUTION]: The skill relies on the execution of several local Python scripts. While the logic within
scripts/composite_layers.pyandscripts/build_prompt.pyappears focused on image blending and text formatting, providing executable scripts as part of the skill increases the attack surface if the environment lacks proper sandboxing.
Audit Metadata