The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its operational model.
Ingestion points: The skill is designed to process user-supplied component requirements, architectural constraints, and design tokens in SKILL.md.
Boundary markers: Absent. There are no instructions for the agent to distinguish between its core logic and instructions that might be embedded within user-provided data.
Capability inventory: The skill is granted access to Bash, WebFetch, and Read tools, which could be misused if the agent is tricked into executing commands hidden in user designs.
Sanitization: None provided. The skill lacks a framework for validating or escaping user-provided input before it influences tool execution or code generation.
[COMMAND_EXECUTION]: The skill requests access to the Bash tool. Although no malicious scripts are present in the provided files, the combination of shell access with a design-generation workflow increases the risk of command injection if the agent interprets user input as instructions.
[EXTERNAL_DOWNLOADS]: The skill references numerous trusted and well-known resources, including official documentation for React, Vue, Svelte, and WCAG standards. These references are used legitimately to provide authoritative guidance for component development.

moai-component-designer