The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Shell tool to execute terminal commands as part of its development workflow for UdonSharp compilation and VRChat SDK operations.- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests untrusted data from project files while maintaining powerful capabilities. 1. Ingestion points: Reads source code and configuration files from the vrchat-project directory (e.g., Assets/, Packages/). 2. Boundary markers: There are no specific delimiters or instructions to treat data from the project files as untrusted content. 3. Capability inventory: The skill utilizes Shell, Write, and StrReplace tools to interact with the filesystem and execute commands. 4. Sanitization: No sanitization or validation protocols are established for handling content retrieved from the codebase.- [DYNAMIC_EXECUTION]: The skill automates the writing and compilation of UdonSharp scripts, which involves runtime code generation for VRChat world interactivity.- [EXTERNAL_DOWNLOADS]: The skill references external documentation from docs.vrchat.com and udosharp-docs.vercel.app, as well as a repository on GitHub for installation.- [REMOTE_CODE_EXECUTION]: The skill includes an instruction to install an external skill from a GitHub repository.

vrchat-dev