The Agent Skills Directory

[COMMAND_EXECUTION]: Modifies system shell configuration files (~/.zshrc, ~/.bash_profile) to establish environment variables and persistent aliases, ensuring the development environment is correctly configured across sessions.\n
Evidence: SKILL.md and references/macos.md contain commands to append configuration lines to shell profile files using redirection and 'sed'.\n- [COMMAND_EXECUTION]: Recommends or executes commands to bypass PowerShell execution policies on Windows systems, which is required to allow the execution of the environment setup scripts.\n
Evidence: references/windows.md includes the command 'Set-ExecutionPolicy Bypass -Scope Process -Force'.\n- [REMOTE_CODE_EXECUTION]: Downloads and executes an installation script from the official NVM (Node Version Manager) repository using a curl-to-bash pipe.\n
Evidence: SKILL.md and references/macos.md contain the command 'curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.40.1/install.sh | bash'.\n- [EXTERNAL_DOWNLOADS]: Fetches the FDK CLI package and specialized Homebrew tap from official vendor-controlled sources (freshdev.io and GitHub).\n
Evidence: SKILL.md references 'https://cdn.freshdev.io/fdk/latest.tgz' and 'brew tap freshworks-developers/homebrew-tap'.\n- [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by interpolating raw user arguments into shell commands that are then processed by autonomous subagents.\n
Ingestion points: The $ARGUMENTS variable in SKILL.md captures user-provided version numbers.\n
Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the subagent prompts.\n
Capability inventory: The skill has 'shell' tool access via the Task subagent system.\n
Sanitization: No sanitization is performed on user arguments beyond a basic whitespace split, allowing potential command fragments to be interpolated into executable strings.

fdk-setup