The Agent Skills Directory

[SAFE]: The skill implements standard project management and development setup procedures without any detected malicious intent or obfuscation.
[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands to configure the Bun runtime environment, specifically writing a bunfig.toml file to the user's configuration directory. This command is used to set a minimumReleaseAge for packages, which is a recognized security control to prevent the installation of recently published (potentially malicious) package versions.
[EXTERNAL_DOWNLOADS]: The skill recommends using official and widely recognized project scaffolding tools via bunx (e.g., create-next-app, create-vite, create-t3-app, create-remix). These downloads originate from well-known registry services and are standard in modern web development.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface where it ingests data from external subagent research and local reference files to generate a project plan. While it lacks explicit boundary markers (e.g., XML tags) for external data, the instructions prioritize synthesizing findings rather than direct execution of untrusted input. The overall capabilities are scoped to project initialization and configuration.

project-planning