The Agent Skills Directory

[SAFE]: The skill defines a structured way for the agent to use a platform-provided tool (poster_render) to generate images from React code. The instructions focus on design aesthetics and technical constraints of the renderer.
[DYNAMIC_EXECUTION]: The skill involves the dynamic generation of TSX code based on user requests, which is then passed to a rendering tool. This script generation is the core intended purpose of the skill and follows a template-driven approach, which is considered standard for this use case.
[INDIRECT_PROMPT_INJECTION]: The skill creates an attack surface by interpolating user-provided data (e.g., names, labels, statistics) into the generated TSX.
Ingestion points: User-provided content for charts, dashboards, and reports is incorporated into the tsx parameter of the poster_render tool.
Boundary markers: The instructions do not explicitly provide delimiters or warnings for the agent to ignore instructions embedded within the user data.
Capability inventory: The skill uses the poster_render tool, which writes files to the local system (e.g., ./chart.png).
Sanitization: No specific sanitization or escaping mechanisms are mandated in the skill instructions, though standard React JSX behavior provides inherent protection against basic script injection.

poster