The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill documentation describes the use of moda canvas edit and <generate> tags, which allow for the execution of JavaScript to automate canvas updates. This execution is performed within a restricted sandbox that blocks high-risk operations such as network access (fetch), dynamic code evaluation (eval), and module loading (require), effectively mitigating common execution-related risks.
[INDIRECT_PROMPT_INJECTION]: The skill accepts user-supplied text to be rendered into images and carousels, representing a vulnerability surface for indirect instructions.
Ingestion points: User-provided topics and descriptions (SKILL.md).
Boundary markers: The skill uses structured recipes and XML tags to delimit content.
Capability inventory: Includes specialized design operations like moda canvas markup, moda canvas edit, and moda media generate-image.
Sanitization: Content is processed by an XML parser and a sandboxed JavaScript environment that restricts access to system resources and network operations.
[COMMAND_EXECUTION]: The skill utilizes a specific set of tools under the moda:* namespace via the Bash tool. These tools are scoped to the intended functionality of the agent and do not involve arbitrary command execution or shell injection vulnerabilities.
[SAFE]: The skill operates within a specialized design toolchain. All referenced tools and commands are legitimate vendor resources provided by the author to facilitate design tasks.

moda-social-instagram