The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill's metadata contains an installation command that retrieves the skill's instructions from a remote GitHub repository. This is a standard procedure for skill management and distribution.
[COMMAND_EXECUTION]: The skill provides the user with a set of shell commands to initialize the project, including virtual environment creation and dependency installation. These instructions are presented as guidance for the end-user rather than commands for the agent to execute autonomously.
[INDIRECT_PROMPT_INJECTION]: The skill defines a process for creating files and directories based on user-supplied project names and paths, which represents a potential surface for indirect prompt injection.
Ingestion points: User-provided project name and target directory (SKILL.md).
Boundary markers: The instructions do not specify any delimiters or safety markers for the user input.
Capability inventory: The agent is tasked with file system operations, including directory creation and file writing.
Sanitization: No specific input validation or sanitization steps are included in the skill's instructions.
[CREDENTIALS_UNSAFE]: The skill promotes secure credential management by advising the user to store the OpenAI API key in a .env file instead of embedding it directly in the backend code.

fullstack-template-generator