The Agent Skills Directory

[PROMPT_INJECTION]: The skill employs strong directive language intended to override the base model's default behaviors. Phrases such as "CRITICAL OUTPUT RULE", "HARD OUTPUT RULE
READ FIRST", and the explicit statement "This rule overrides any model default that wants to collapse output into a single image" are used to force the model to adhere to the skill's specific multi-image formatting requirements.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external, untrusted user data to drive its image generation logic.
Ingestion points: The skill dynamically interprets user design briefs (e.g., industry, brand mood, specific style requests) to adjust its "Combinatorial Variation Engine".
Boundary markers: No specific delimiters or instructions to ignore embedded commands within the user brief are present in the documentation.
Capability inventory: The skill leverages image generation tools to produce visual design references based on user input.
Sanitization: There is no evidence of sanitization, validation, or escaping logic applied to user-supplied text before it influences the generation prompts.
[NO_CODE]: The analyzed skill consists entirely of instructional Markdown content and YAML metadata; it does not include or reference any executable scripts, binary files, or external software dependencies.

imagegen-frontend-web