The Agent Skills Directory

[SAFE]: The skill is a legitimate design extension. All tool calls and instructions are consistent with the stated purpose of generating video covers and photo-mode posts. The use of the platform-native moda tools indicates a restricted and safe execution environment.
[COMMAND_EXECUTION]: The skill utilizes a set of domain-specific commands (e.g., moda canvas, moda media, moda export) via a restricted Bash tool. These operations are limited to design asset management and do not pose a risk of arbitrary system command execution.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided topics and intents to construct prompts for image generation and canvas metadata. While this represents a standard ingestion surface, the risk is mitigated by the scoped nature of the resulting tool operations.
Ingestion points: User-supplied [Topic] and [intent] arguments in the SKILL.md recipes.
Boundary markers: None explicitly defined; content is interpolated into tool arguments.
Capability inventory: Access is restricted to moda ecosystem tools for design and media production.
Sanitization: Relies on the underlying platform's handling of CLI arguments and LLM safety layers.

moda-social-tiktok