The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes several command-line tools to facilitate the development and deployment process.\n
It utilizes joyco create and joyco lab, which are specialized tools provided by the vendor joyco-studio for experiment management.\n
It runs vercel --prod to deploy the isolated experiments to the Vercel hosting platform.\n
It executes npm run dev to start a local development server for the ported experiment code.\n- [EXTERNAL_DOWNLOADS]: The skill performs dynamic package installation using npm install.\n
The packages to be installed are determined by parsing imports from user-provided code, which creates a risk of dependency confusion or typosquatting if the user's source code contains malicious imports.\n- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes untrusted user data to drive its automation.\n
Ingestion points: The agent reads and interprets user-provided code blocks to extract functional descriptions, identify required templates, and list dependencies.\n
Boundary markers: The skill does not implement delimiters or instructions to ignore commands that may be embedded within the user's code.\n
Capability inventory: The skill possesses broad capabilities, including the ability to execute shell commands, install software, and deploy to external web services.\n
Sanitization: There is no evidence of validation or sanitization of the content extracted from the user code before it is passed to CLI tools or used to generate registry entries.

joyco-lab