The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection as it is designed to ingest and process untrusted external data from various sources.
Ingestion points: The skill processes external data via commands like /design-review <file or screenshot>, /figma <URL>, /ux-audit <brief>, and /site-to-figma (SKILL.md, references/figma-workflow.md).
Boundary markers: There are no explicit instructions or delimiters defined to separate untrusted content from the agent's instructions, nor are there warnings to ignore potentially embedded commands in the processed files.
Capability inventory: The skill has access to powerful capabilities including file system writing (token generation), network operations (Firebase Hosting, preview server), and subprocess execution (npm, npx) (references/deployment.md).
Sanitization: No explicit sanitization or validation of the content extracted from Figma URLs or external HTML files is mentioned in the instruction set.
[COMMAND_EXECUTION]: The skill facilitates the execution of local commands for development workflows.
It uses npx serve and npm run dev to start local preview servers for design verification (references/deployment.md).
It provides a troubleshooting command using lsof, grep, and kill to manage occupied local ports used by the Figma Desktop Bridge (references/figma-creation.md).
[EXTERNAL_DOWNLOADS]: The skill references and downloads assets from several well-known and trusted services.
Typography: Connects to fonts.googleapis.com and fonts.gstatic.com for font loading (references/deployment.md).
Icons: Fetches Lucide icons via CDN (SKILL.md).
Tools: Integrates with official Figma and Firebase platforms for design extraction and site deployment.

design